feat(abonnements): paliers LLM Essentielle/Premium, admin modeles par forfait, statistiques revenus, emails marketing

- Gamme officielle : Essentielle (Pro) = deepseek-v4-flash, glm-5.3-flash, minimax-m3 ; Premium (Business) = claude-sonnet-5, deepseek-v4-pro, glm-5.3 - Routage reel : le modele employe suit le palier IA du forfait (reglages admin > defaut du plan), garde anti-croisement de palier - Nouvelle page admin « Modeles & abonnements » (matrice forfaits/paliers, modele par defaut, catalogue OpenRouter) - Statistiques enrichies : revenus (30 j + total), MRR estime, credits, liste d'attente, paliers utilises - Nouvelle page admin « Marketing » : audiences avec compteurs, apercu, envoi test obligatoire, historique, desabonnement public + en-tetes List-Unsubscribe ; .gitignore pour les donnees d'execution - Interface (accueil + tarifs) alignee sur la gamme, 13 langues completes - Tests : routage par palier (fonction + tache), marketing, revenus en base
2026-09-05 12:51:49 +02:00
parent 4e6850d6b1
commit 6c26712687
76 changed files with 5289 additions and 271 deletions
--- a/frontend/src/app/admin/RevenueOverview.tsx
+++ b/frontend/src/app/admin/RevenueOverview.tsx
@@ -0,0 +1,159 @@
+"use client";
+
+import { useEffect, useState } from "react";
+import { Euro, TrendingUp, Users, Cpu, Loader2 } from "lucide-react";
+import { Card, CardContent, CardHeader, CardTitle } from "@/components/ui/card";
+import { useTranslationStore } from "@/lib/store";
+import { API_BASE } from "@/lib/config";
+import { useI18n } from "@/lib/i18n";
+import type { AdminStatsResponse } from "./types";
+
+const getToken = () => useTranslationStore.getState().settings.adminToken ?? "";
+
+function formatMoney(value: number, currency: string): string {
+  const currencyLabel = currency === "EUR" ? "€" : currency === "USD" ? "$" : currency;
+  return `${value.toLocaleString("fr-FR", { maximumFractionDigits: 2 })} ${currencyLabel}`;
+}
+
+/**
+ * Cartes business (revenus, MRR estimé, crédits, liste d'attente, paliers IA)
+ * alimentées par GET /api/v1/admin/stats (champs revenue, mrr_estimated,
+ * waitlist_count, ai_tier_usage).
+ */
+export function RevenueOverview() {
+  const { t } = useI18n();
+  const [stats, setStats] = useState<AdminStatsResponse | null>(null);
+  const [isLoading, setIsLoading] = useState(true);
+  const [error, setError] = useState<string | null>(null);
+
+  useEffect(() => {
+    let cancelled = false;
+    const load = async () => {
+      try {
+        const response = await fetch(`${API_BASE}/api/v1/admin/stats`, {
+          headers: { Authorization: `Bearer ${getToken()}` },
+        });
+        if (!response.ok) throw new Error(`HTTP ${response.status}`);
+        const json = await response.json();
+        if (!cancelled) {
+          setStats(json.data ?? json);
+          setError(null);
+        }
+      } catch (e) {
+        if (!cancelled) setError(e instanceof Error ? e.message : "Erreur");
+      } finally {
+        if (!cancelled) setIsLoading(false);
+      }
+    };
+    load();
+    const interval = setInterval(load, 30000);
+    return () => {
+      cancelled = true;
+      clearInterval(interval);
+    };
+  }, []);
+
+  if (isLoading && !stats) {
+    return (
+      <div className="grid gap-4 md:grid-cols-2 lg:grid-cols-5">
+        {Array.from({ length: 5 }).map((_, i) => (
+          <Card key={i}>
+            <CardHeader className="flex flex-row items-center justify-between space-y-0 pb-2">
+              <div className="h-4 w-24 animate-pulse rounded bg-muted" />
+            </CardHeader>
+            <CardContent>
+              <div className="h-8 w-20 animate-pulse rounded bg-muted" />
+            </CardContent>
+          </Card>
+        ))}
+      </div>
+    );
+  }
+
+  if (error || !stats) {
+    return (
+      <Card>
+        <CardContent className="flex items-center gap-2 py-4 text-sm text-red-500">
+          <Cpu className="size-4" /> {t("admin.stats.unavailable", { error: error ?? "" })}
+        </CardContent>
+      </Card>
+    );
+  }
+
+  const currency = stats.revenue?.currency ?? "EUR";
+  const tierTotal =
+    (stats.ai_tier_usage?.essential ?? 0) +
+    (stats.ai_tier_usage?.premium ?? 0) +
+    (stats.ai_tier_usage?.classic ?? 0) +
+    (stats.ai_tier_usage?.other ?? 0);
+
+  const cards = [
+    {
+      title: t("admin.stats.revenue30"),
+      icon: <Euro className="h-4 w-4 text-muted-foreground" />,
+      value: formatMoney(stats.revenue?.collected_30d ?? 0, currency),
+      sub: t("admin.stats.payments30", { count: stats.revenue?.payments_30d ?? 0 }),
+    },
+    {
+      title: t("admin.stats.revenueTotal"),
+      icon: <TrendingUp className="h-4 w-4 text-muted-foreground" />,
+      value: formatMoney(stats.revenue?.collected_total ?? 0, currency),
+      sub: t("admin.stats.includingCredits", {
+        amount: formatMoney(stats.revenue?.credits_purchased ?? 0, currency),
+      }),
+    },
+    {
+      title: t("admin.stats.mrr"),
+      icon: <TrendingUp className="h-4 w-4 text-muted-foreground" />,
+      value: formatMoney(stats.mrr_estimated?.total ?? 0, currency),
+      sub:
+        Object.entries(stats.mrr_estimated?.by_plan ?? {})
+          .map(([plan, mrr]) => `${plan} : ${formatMoney(mrr, currency)}`)
+          .join(" · ") || "—",
+    },
+    {
+      title: t("admin.stats.waitlist"),
+      icon: <Users className="h-4 w-4 text-muted-foreground" />,
+      value: String(stats.waitlist_count ?? 0),
+      sub: t("admin.stats.waitlistSub"),
+    },
+    {
+      title: t("admin.stats.aiTiers"),
+      icon: <Cpu className="h-4 w-4 text-muted-foreground" />,
+      value: `${stats.ai_tier_usage?.essential ?? 0} / ${stats.ai_tier_usage?.premium ?? 0}`,
+      sub: t("admin.stats.tiersSub", {
+        classic: stats.ai_tier_usage?.classic ?? 0,
+        other: stats.ai_tier_usage?.other ?? 0,
+      }),
+    },
+  ];
+
+  return (
+    <div>
+      <div className="grid gap-4 md:grid-cols-2 lg:grid-cols-5">
+        {cards.map((c) => (
+          <Card key={c.title}>
+            <CardHeader className="flex flex-row items-center justify-between space-y-0 pb-2">
+              <CardTitle className="text-sm font-medium">{c.title}</CardTitle>
+              {c.icon}
+            </CardHeader>
+            <CardContent>
+              <div className="text-2xl font-bold">{c.value}</div>
+              <p className="truncate text-xs text-muted-foreground" title={c.sub}>
+                {c.sub}
+              </p>
+            </CardContent>
+          </Card>
+        ))}
+      </div>
+      {tierTotal === 0 && (
+        <p className="mt-2 text-xs text-muted-foreground">{t("admin.stats.noTranslationsYet")}</p>
+      )}
+      {isLoading && (
+        <p className="mt-2 flex items-center gap-1.5 text-xs text-muted-foreground">
+          <Loader2 className="size-3 animate-spin" /> {t("admin.stats.refreshing")}
+        </p>
+      )}
+    </div>
+  );
+}
--- a/frontend/src/app/admin/constants.ts
+++ b/frontend/src/app/admin/constants.ts
@@ -1,4 +1,4 @@
-import { CreditCard, LayoutDashboard, Settings, FileText, Users, type LucideIcon } from 'lucide-react';
+import { CreditCard, LayoutDashboard, Settings, FileText, Users, Cpu, Megaphone, BarChart3, type LucideIcon } from 'lucide-react';

 export interface AdminNavItem {
  labelKey: string;
@@ -8,9 +8,12 @@ export interface AdminNavItem {

 export const adminNavItems: AdminNavItem[] = [
  { labelKey: 'admin.nav.dashboard', href: '/admin', icon: LayoutDashboard },
+  { labelKey: 'admin.nav.stats', href: '/admin/stats', icon: BarChart3 },
  { labelKey: 'admin.nav.users', href: '/admin/users', icon: Users },
  { labelKey: 'admin.nav.pricing', href: '/admin/pricing', icon: CreditCard },
+  { labelKey: 'admin.nav.models', href: '/admin/models', icon: Cpu },
  { labelKey: 'admin.nav.providers', href: '/admin/settings', icon: Settings },
+  { labelKey: 'admin.nav.marketing', href: '/admin/marketing', icon: Megaphone },
  { labelKey: 'admin.nav.system', href: '/admin/system', icon: Settings },
  { labelKey: 'admin.nav.logs', href: '/admin/logs', icon: FileText },
 ];
--- a/frontend/src/app/admin/marketing/page.tsx
+++ b/frontend/src/app/admin/marketing/page.tsx
@@ -0,0 +1,426 @@
+"use client";
+
+import { useState, useEffect, useMemo } from "react";
+import {
+  Megaphone,
+  Loader2,
+  Send,
+  FlaskConical,
+  History,
+  AlertCircle,
+  CheckCircle2,
+  Eye,
+} from "lucide-react";
+import { Card, CardContent, CardDescription, CardHeader, CardTitle } from "@/components/ui/card";
+import { Button } from "@/components/ui/button";
+import { Input } from "@/components/ui/input";
+import { Label } from "@/components/ui/label";
+import { Textarea } from "@/components/ui/textarea";
+import { Badge } from "@/components/ui/badge";
+import { Select, SelectContent, SelectItem, SelectTrigger, SelectValue } from "@/components/ui/select";
+import { useNotification } from "@/components/ui/notification";
+import { useTranslationStore } from "@/lib/store";
+import { API_BASE } from "@/lib/config";
+import { useI18n } from "@/lib/i18n";
+import type {
+  MarketingAudiencesResponse,
+  MarketingEmailHistoryEntry,
+  MarketingEmailSendResponse,
+} from "../types";
+
+const getToken = () => useTranslationStore.getState().settings.adminToken ?? "";
+
+/** Clés i18n par identifiant d'audience (l'identifiant part en base/historique). */
+const AUDIENCE_LABEL_KEYS: Record<string, string> = {
+  waitlist: "admin.marketing.audienceWaitlist",
+  inactive_30: "admin.marketing.audienceInactive",
+  "plan:pro": "admin.marketing.audiencePlanPro",
+  "plan:starter": "admin.marketing.audiencePlanStarter",
+  "plan:business": "admin.marketing.audiencePlanBusiness",
+  "plan:enterprise": "admin.marketing.audiencePlanEnterprise",
+  "plan:free": "admin.marketing.audiencePlanFree",
+  all_users: "admin.marketing.audienceAllUsers",
+};
+
+export default function AdminMarketingPage() {
+  const { t } = useI18n();
+  const { success, error } = useNotification();
+  const [audiences, setAudiences] = useState<Record<string, number>>({});
+  const [unsubscribedCount, setUnsubscribedCount] = useState(0);
+  const [audience, setAudience] = useState("waitlist");
+  const [subject, setSubject] = useState("");
+  const [html, setHtml] = useState("");
+  const [testEmail, setTestEmail] = useState("");
+  const [showPreview, setShowPreview] = useState(false);
+  const [isLoading, setIsLoading] = useState(true);
+  const [isSendingTest, setIsSendingTest] = useState(false);
+  const [isSending, setIsSending] = useState(false);
+  const [history, setHistory] = useState<MarketingEmailHistoryEntry[]>([]);
+  /** Empreinte SHA-256 du contenu courant (même calcul que le backend). */
+  const [contentHash, setContentHash] = useState<string | null>(null);
+  /** false si le navigateur n'expose pas crypto.subtle (indicateur absent,
+   * l'envoi reste possible : le serveur valide l'envoi test). */
+  const [hashSupported, setHashSupported] = useState(true);
+
+  useEffect(() => {
+    if (!crypto?.subtle) {
+      setHashSupported(false);
+      setContentHash(null);
+      return;
+    }
+    let cancelled = false;
+    const data = new TextEncoder().encode(`${subject.trim()}\n${html.trim()}`);
+    crypto.subtle
+      .digest("SHA-256", data)
+      .then((buf) => {
+        if (cancelled) return;
+        const hex = Array.from(new Uint8Array(buf))
+          .map((b) => b.toString(16).padStart(2, "0"))
+          .join("");
+        setContentHash(hex);
+      })
+      .catch(() => setContentHash(null));
+    return () => {
+      cancelled = true;
+    };
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [subject, html]);
+
+  const loadAudiences = async () => {
+    const response = await fetch(`${API_BASE}/api/v1/admin/marketing/audiences`, {
+      headers: { Authorization: `Bearer ${getToken()}` },
+    });
+    if (!response.ok) throw new Error(`HTTP ${response.status}`);
+    const json: MarketingAudiencesResponse = await response.json();
+    setAudiences(json.data.audiences);
+    setUnsubscribedCount(json.data.unsubscribed_count);
+  };
+
+  const loadHistory = async () => {
+    const response = await fetch(`${API_BASE}/api/v1/admin/marketing/email/history`, {
+      headers: { Authorization: `Bearer ${getToken()}` },
+    });
+    if (!response.ok) throw new Error(`HTTP ${response.status}`);
+    const json = await response.json();
+    setHistory(json.data ?? []);
+  };
+
+  useEffect(() => {
+    setIsLoading(true);
+    Promise.all([loadAudiences(), loadHistory()])
+      .catch((e) =>
+        error({ title: t("admin.marketing.loadErrorTitle"), description: String(e.message ?? e) })
+      )
+      .finally(() => setIsLoading(false));
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, []);
+
+  const audienceLabel = (key: string): string => {
+    const labelKey = AUDIENCE_LABEL_KEYS[key];
+    return labelKey ? t(labelKey) : key;
+  };
+
+  const recipientCount = audiences[audience] ?? 0;
+
+  /** Simple indicateur (non bloquant) : un test réussi de ce contenu exact
+   * apparaît dans l'historique chargé. La décision d'autoriser l'envoi réel
+   * appartient au serveur (l'historique affiché est tronqué à 100 entrées). */
+  const testSeenInHistory =
+    !!contentHash &&
+    history.some(
+      (h) => h.test_mode && h.sent_count > 0 && h.content_hash === contentHash
+    );
+
+  const sendTest = async () => {
+    if (!subject.trim() || !html.trim()) {
+      error({ title: t("admin.marketing.incompleteTitle"), description: t("admin.marketing.incompleteDesc") });
+      return;
+    }
+    setIsSendingTest(true);
+    try {
+      const response = await fetch(`${API_BASE}/api/v1/admin/marketing/email/send`, {
+        method: "POST",
+        headers: {
+          Authorization: `Bearer ${getToken()}`,
+          "Content-Type": "application/json",
+        },
+        body: JSON.stringify({
+          audience,
+          subject: subject.trim(),
+          html: html.trim(),
+          test_mode: true,
+          test_email: testEmail.trim() || null,
+        }),
+      });
+      const json = await response.json().catch(() => ({}));
+      if (response.ok) {
+        const d: MarketingEmailSendResponse["data"] = json.data;
+        success({
+          title: t("admin.marketing.testDoneTitle"),
+          description: t("admin.marketing.testDoneDesc", { email: d.recipient ?? "" }),
+        });
+        await loadHistory();
+      } else {
+        error({
+          title: t("admin.marketing.testFailedTitle"),
+          description: json.message || `HTTP ${response.status}`,
+        });
+        await loadHistory();
+      }
+    } catch {
+      error({ title: t("admin.marketing.loadErrorTitle"), description: t("admin.marketing.networkDesc") });
+    } finally {
+      setIsSendingTest(false);
+    }
+  };
+
+  const sendReal = async () => {
+    if (!subject.trim() || !html.trim()) {
+      error({ title: t("admin.marketing.incompleteTitle"), description: t("admin.marketing.incompleteDesc") });
+      return;
+    }
+    if (
+      !window.confirm(
+        t("admin.marketing.confirmSend", { count: recipientCount, audience: audienceLabel(audience) })
+      )
+    ) {
+      return;
+    }
+    setIsSending(true);
+    try {
+      const response = await fetch(`${API_BASE}/api/v1/admin/marketing/email/send`, {
+        method: "POST",
+        headers: {
+          Authorization: `Bearer ${getToken()}`,
+          "Content-Type": "application/json",
+        },
+        body: JSON.stringify({
+          audience,
+          subject: subject.trim(),
+          html: html.trim(),
+          test_mode: false,
+        }),
+      });
+      const json = await response.json().catch(() => ({}));
+      if (response.ok) {
+        const d: MarketingEmailSendResponse["data"] = json.data;
+        success({
+          title: t("admin.marketing.sendingTitle"),
+          description: t("admin.marketing.sendingDesc", { count: d.queued ?? 0 }),
+        });
+        setTimeout(() => {
+          loadHistory().catch(() => {});
+        }, 3000);
+      } else {
+        error({
+          title: t("admin.marketing.sendRefusedTitle"),
+          description: json.message || `HTTP ${response.status}`,
+        });
+      }
+    } catch {
+      error({ title: t("admin.marketing.loadErrorTitle"), description: t("admin.marketing.networkDesc") });
+    } finally {
+      setIsSending(false);
+    }
+  };
+
+  const previewDoc = useMemo(
+    () =>
+      html.trim() ||
+      `<html><body style='font-family:sans-serif;color:#94a3b8;padding:32px;text-align:center;'>${t(
+        "admin.marketing.previewEmpty"
+      )}</body></html>`,
+    [html, t]
+  );
+
+  return (
+    <div className="space-y-6">
+      <div className="flex items-center gap-3">
+        <div className="w-10 h-10 bg-purple-600/20 rounded-lg flex items-center justify-center">
+          <Megaphone className="w-5 h-5 text-purple-400" />
+        </div>
+        <div>
+          <h1 className="text-xl font-semibold text-foreground">{t("admin.marketing.title")}</h1>
+          <p className="text-sm text-muted-foreground">{t("admin.marketing.subtitle")}</p>
+        </div>
+      </div>
+
+      {isLoading ? (
+        <div className="flex items-center justify-center py-12">
+          <Loader2 className="size-8 animate-spin text-muted-foreground" />
+        </div>
+      ) : (
+        <div className="grid gap-6 lg:grid-cols-3">
+          {/* ── Composition ── */}
+          <Card className="lg:col-span-2 overflow-visible">
+            <CardHeader>
+              <CardTitle className="text-base">{t("admin.marketing.newCampaign")}</CardTitle>
+              <CardDescription>
+                {t("admin.marketing.unsubscribedNote", { count: unsubscribedCount })}
+              </CardDescription>
+            </CardHeader>
+            <CardContent className="space-y-4">
+              <div className="grid gap-4 sm:grid-cols-2">
+                <div className="space-y-2">
+                  <Label>{t("admin.marketing.audience")}</Label>
+                  <Select value={audience} onValueChange={setAudience}>
+                    <SelectTrigger>
+                      <SelectValue placeholder={t("admin.marketing.audiencePlaceholder")} />
+                    </SelectTrigger>
+                    <SelectContent>
+                      {Object.keys(audiences)
+                        .sort((a, b) => (audiences[b] ?? 0) - (audiences[a] ?? 0))
+                        .map((key) => (
+                          <SelectItem key={key} value={key}>
+                            {audienceLabel(key)} ({audiences[key]})
+                          </SelectItem>
+                        ))}
+                    </SelectContent>
+                  </Select>
+                  <p className="text-xs text-muted-foreground">
+                    {t("admin.marketing.recipientsAvailable", { count: recipientCount })}
+                  </p>
+                </div>
+                <div className="space-y-2">
+                  <Label htmlFor="marketing-test-email">{t("admin.marketing.testEmail")}</Label>
+                  <Input
+                    id="marketing-test-email"
+                    type="email"
+                    placeholder={t("admin.marketing.testEmailPlaceholder")}
+                    value={testEmail}
+                    onChange={(e) => setTestEmail(e.target.value)}
+                  />
+                </div>
+              </div>
+
+              <div className="space-y-2">
+                <Label htmlFor="marketing-subject">{t("admin.marketing.subject")}</Label>
+                <Input
+                  id="marketing-subject"
+                  placeholder={t("admin.marketing.subjectPlaceholder")}
+                  value={subject}
+                  onChange={(e) => setSubject(e.target.value)}
+                  maxLength={200}
+                />
+              </div>
+
+              <div className="space-y-2">
+                <div className="flex items-center justify-between">
+                  <Label htmlFor="marketing-html">{t("admin.marketing.html")}</Label>
+                  <Button
+                    variant="ghost"
+                    size="sm"
+                    className="h-7 gap-1.5 text-xs"
+                    onClick={() => setShowPreview((v) => !v)}
+                  >
+                    <Eye className="size-3" />
+                    {showPreview ? t("admin.marketing.hidePreview") : t("admin.marketing.preview")}
+                  </Button>
+                </div>
+                <Textarea
+                  id="marketing-html"
+                  className="min-h-48 font-mono text-xs"
+                  placeholder={t("admin.marketing.htmlPlaceholder")}
+                  value={html}
+                  onChange={(e) => setHtml(e.target.value)}
+                />
+                <p className="text-xs text-muted-foreground">{t("admin.marketing.footerNote")}</p>
+                {showPreview && (
+                  <iframe
+                    title={t("admin.marketing.previewTitle")}
+                    srcDoc={previewDoc}
+                    className="h-72 w-full rounded-lg border border-border bg-white"
+                    sandbox=""
+                  />
+                )}
+              </div>
+
+              <div className="flex flex-wrap items-center gap-3 border-t border-border/60 pt-4">
+                <Button variant="outline" onClick={sendTest} disabled={isSendingTest} className="gap-2">
+                  {isSendingTest ? (
+                    <Loader2 className="size-4 animate-spin" />
+                  ) : (
+                    <FlaskConical className="size-4" />
+                  )}
+                  {t("admin.marketing.sendTest")}
+                </Button>
+                <Button
+                  onClick={sendReal}
+                  disabled={isSending || recipientCount === 0}
+                  className="gap-2"
+                >
+                  {isSending ? (
+                    <Loader2 className="size-4 animate-spin" />
+                  ) : (
+                    <Send className="size-4" />
+                  )}
+                  {t("admin.marketing.sendReal", { count: recipientCount })}
+                </Button>
+                {!hashSupported && (
+                  <span className="flex items-center gap-1.5 text-xs text-muted-foreground">
+                    <AlertCircle className="size-3.5" />
+                    {t("admin.marketing.localCheckUnavailable")}
+                  </span>
+                )}
+                {hashSupported && testSeenInHistory && (
+                  <span className="flex items-center gap-1.5 text-xs text-green-500">
+                    <CheckCircle2 className="size-3.5" />
+                    {t("admin.marketing.testValidated")}
+                  </span>
+                )}
+                {hashSupported && !testSeenInHistory && (
+                  <span className="flex items-center gap-1.5 text-xs text-amber-500">
+                    <AlertCircle className="size-3.5" />
+                    {t("admin.marketing.testRequiredHint")}
+                  </span>
+                )}
+              </div>
+            </CardContent>
+          </Card>
+
+          {/* ── Historique ── */}
+          <Card>
+            <CardHeader className="flex flex-row items-center gap-2 pb-3">
+              <History className="size-4 text-muted-foreground" />
+              <CardTitle className="text-base">{t("admin.marketing.history")}</CardTitle>
+            </CardHeader>
+            <CardContent className="space-y-3">
+              {history.length === 0 && (
+                <p className="text-sm text-muted-foreground">{t("admin.marketing.noHistory")}</p>
+              )}
+              {history.map((h) => (
+                <div key={h.id} className="rounded-lg border border-border/60 p-3 text-xs">
+                  <div className="mb-1 flex flex-wrap items-center gap-2">
+                    {h.test_mode ? (
+                      <Badge variant="outline" className="text-[10px]">
+                        {t("admin.marketing.badgeTest")}
+                      </Badge>
+                    ) : (
+                      <Badge className="bg-accent/10 text-accent border-accent/30 text-[10px]">
+                        {t("admin.marketing.badgeReal")}
+                      </Badge>
+                    )}
+                    <span className="text-muted-foreground">
+                      {new Date(h.sent_at).toLocaleString("fr-FR")}
+                    </span>
+                  </div>
+                  <p className="mb-1 font-medium text-foreground">{h.subject}</p>
+                  <p className="text-muted-foreground">
+                    {audienceLabel(h.audience)} —{" "}
+                    {t("admin.marketing.historySent", {
+                      sent: h.sent_count,
+                      total: h.recipients_total,
+                    })}
+                    {h.failed_count > 0
+                      ? t("admin.marketing.historyFailedSuffix", { count: h.failed_count })
+                      : ""}
+                  </p>
+                </div>
+              ))}
+            </CardContent>
+          </Card>
+        </div>
+      )}
+    </div>
+  );
+}
--- a/frontend/src/app/admin/models/page.tsx
+++ b/frontend/src/app/admin/models/page.tsx
@@ -0,0 +1,375 @@
+"use client";
+
+import { useState, useEffect } from "react";
+import { Cpu, Save, Loader2, ArrowUp, ArrowDown, X, Plus, RefreshCw } from "lucide-react";
+import { Card, CardContent, CardDescription, CardHeader, CardTitle } from "@/components/ui/card";
+import { Button } from "@/components/ui/button";
+import { Badge } from "@/components/ui/badge";
+import { useNotification } from "@/components/ui/notification";
+import { ModelCombobox, ModelOption } from "@/components/ui/model-combobox";
+import { useTranslationStore } from "@/lib/store";
+import { API_BASE } from "@/lib/config";
+import { useI18n } from "@/lib/i18n";
+import type { AiTierSettings, AiTiersSettings } from "../types";
+
+interface FullSettings {
+  [key: string]: unknown;
+}
+
+const getToken = () => useTranslationStore.getState().settings.adminToken ?? "";
+
+/** Éditeur d'un palier : liste ordonnée (ordre = priorité de secours) + radio « par défaut ». */
+function TierEditor({
+  tier,
+  onChange,
+  catalog,
+  onLoadCatalog,
+  catalogLoading,
+}: {
+  tier: AiTierSettings;
+  onChange: (next: AiTierSettings) => void;
+  catalog: ModelOption[];
+  onLoadCatalog: () => void;
+  catalogLoading: boolean;
+}) {
+  const { t } = useI18n();
+  const [adding, setAdding] = useState(false);
+
+  const move = (index: number, dir: -1 | 1) => {
+    const models = [...tier.models];
+    const target = index + dir;
+    if (target < 0 || target >= models.length) return;
+    [models[index], models[target]] = [models[target], models[index]];
+    onChange({ ...tier, models });
+  };
+
+  const remove = (index: number) => {
+    const models = tier.models.filter((_, i) => i !== index);
+    onChange({ ...tier, models });
+  };
+
+  const add = (id: string) => {
+    const idClean = id.trim();
+    if (!idClean || tier.models.includes(idClean)) {
+      setAdding(false);
+      return;
+    }
+    onChange({ ...tier, models: [...tier.models, idClean] });
+    setAdding(false);
+  };
+
+  const defaultModel = tier.default_model || tier.models[0] || "";
+
+  return (
+    <div className="space-y-2">
+      {tier.models.map((model, i) => {
+        const isDefault = model === defaultModel;
+        return (
+          <div
+            key={model}
+            className={
+              "flex items-center gap-2 rounded-lg border px-3 py-2 " +
+              (isDefault ? "border-accent/50 bg-accent/5" : "border-border/60 bg-card")
+            }
+          >
+            <button
+              type="button"
+              role="radio"
+              aria-checked={isDefault}
+              aria-label={t("admin.models.setDefault", { model })}
+              onClick={() => onChange({ ...tier, default_model: model })}
+              className={
+                "flex size-4 shrink-0 items-center justify-center rounded-full border " +
+                (isDefault ? "border-accent" : "border-muted-foreground/40")
+              }
+            >
+              {isDefault && <span className="size-2 rounded-full bg-accent" />}
+            </button>
+            <span className="flex-1 cursor-pointer font-mono text-xs" onClick={() => onChange({ ...tier, default_model: model })}>
+              {model}
+              {isDefault ? (
+                <Badge variant="secondary" className="ms-2 font-sans text-[10px]">
+                  {t("admin.models.defaultBadge")}
+                </Badge>
+              ) : i === 0 ? (
+                <span className="ms-2 font-sans text-[10px] font-bold uppercase tracking-wider text-muted-foreground">
+                  {t("admin.models.fallbackFirst")}
+                </span>
+              ) : null}
+            </span>
+            <Button
+              variant="ghost"
+              size="sm"
+              className="h-6 w-6 p-0"
+              onClick={() => move(i, -1)}
+              disabled={i === 0}
+              aria-label={t("admin.models.moveUp")}
+            >
+              <ArrowUp className="size-3" />
+            </Button>
+            <Button
+              variant="ghost"
+              size="sm"
+              className="h-6 w-6 p-0"
+              onClick={() => move(i, 1)}
+              disabled={i === tier.models.length - 1}
+              aria-label={t("admin.models.moveDown")}
+            >
+              <ArrowDown className="size-3" />
+            </Button>
+            <Button
+              variant="ghost"
+              size="sm"
+              className="h-6 w-6 p-0 text-red-500"
+              onClick={() => remove(i)}
+              aria-label={t("admin.models.remove")}
+            >
+              <X className="size-3" />
+            </Button>
+          </div>
+        );
+      })}
+
+      {tier.models.length === 0 && (
+        <p className="text-xs text-muted-foreground">{t("admin.models.emptyTier")}</p>
+      )}
+
+      {adding ? (
+        <div className="space-y-2 rounded-lg border border-dashed border-border p-3">
+          <ModelCombobox
+            value=""
+            onChange={add}
+            models={catalog}
+            isLoading={catalogLoading}
+            onFetchModels={onLoadCatalog}
+            providerLabel="OpenRouter"
+            placeholder="provider/nom-du-modèle"
+          />
+          <Button variant="ghost" size="sm" onClick={() => setAdding(false)}>
+            {t("admin.models.cancel")}
+          </Button>
+        </div>
+      ) : (
+        <div className="flex items-center gap-2 pt-1">
+          <Button variant="outline" size="sm" onClick={() => setAdding(true)} className="gap-1.5">
+            <Plus className="size-3" /> {t("admin.models.addModel")}
+          </Button>
+          <Button
+            variant="ghost"
+            size="sm"
+            onClick={onLoadCatalog}
+            disabled={catalogLoading}
+            className="h-7 gap-1.5 px-2 text-xs text-muted-foreground"
+          >
+            {catalogLoading ? (
+              <Loader2 className="size-3 animate-spin" />
+            ) : (
+              <RefreshCw className="size-3" />
+            )}
+            {t("admin.models.catalog")}
+          </Button>
+        </div>
+      )}
+    </div>
+  );
+}
+
+export default function AdminModelsPage() {
+  const { t } = useI18n();
+  const { success, error } = useNotification();
+  const [aiTiers, setAiTiers] = useState<AiTiersSettings | null>(null);
+  const [fullSettings, setFullSettings] = useState<FullSettings>({});
+  const [isLoading, setIsLoading] = useState(true);
+  const [isSaving, setIsSaving] = useState(false);
+  const [catalog, setCatalog] = useState<ModelOption[]>([]);
+  const [catalogLoading, setCatalogLoading] = useState(false);
+
+  useEffect(() => {
+    loadSettings();
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, []);
+
+  const loadSettings = async () => {
+    setIsLoading(true);
+    try {
+      const response = await fetch(`${API_BASE}/api/v1/admin/settings`, {
+        headers: { Authorization: `Bearer ${getToken()}` },
+      });
+      if (response.ok) {
+        const envelope = await response.json();
+        const payload = envelope.data ?? envelope;
+        setFullSettings(payload);
+        if (payload.ai_tiers) {
+          setAiTiers(payload.ai_tiers);
+        }
+      } else {
+        error({
+          title: t("admin.models.loadErrorTitle"),
+          description: t("admin.models.loadErrorDesc", { status: response.status }),
+        });
+      }
+    } catch {
+      error({ title: t("admin.models.networkTitle"), description: t("admin.models.networkDesc") });
+    } finally {
+      setIsLoading(false);
+    }
+  };
+
+  const loadCatalog = async () => {
+    setCatalogLoading(true);
+    try {
+      const response = await fetch(`${API_BASE}/api/v1/admin/providers/openrouter/models`, {
+        headers: { Authorization: `Bearer ${getToken()}` },
+      });
+      if (response.ok) {
+        const data = await response.json();
+        setCatalog(data.data || []);
+      } else {
+        error({
+          title: t("admin.models.catalogErrorTitle"),
+          description: t("admin.models.loadErrorDesc", { status: response.status }),
+        });
+      }
+    } catch {
+      error({
+        title: t("admin.models.catalogErrorTitle"),
+        description: t("admin.models.catalogErrorDesc"),
+      });
+    } finally {
+      setCatalogLoading(false);
+    }
+  };
+
+  const save = async () => {
+    if (!aiTiers) return;
+    setIsSaving(true);
+    try {
+      const body = { ...fullSettings, ai_tiers: aiTiers };
+      const response = await fetch(`${API_BASE}/api/v1/admin/settings`, {
+        method: "PUT",
+        headers: {
+          Authorization: `Bearer ${getToken()}`,
+          "Content-Type": "application/json",
+        },
+        body: JSON.stringify(body),
+      });
+      if (response.ok) {
+        const envelope = await response.json();
+        const payload = envelope.data ?? envelope;
+        if (payload?.ai_tiers) setAiTiers(payload.ai_tiers);
+        success({
+          title: t("admin.models.savedTitle"),
+          description: t("admin.models.savedDesc"),
+        });
+      } else {
+        const detail = await response.json().catch(() => ({}));
+        error({
+          title: t("admin.models.saveErrorTitle"),
+          description: detail.detail || detail.message || t("admin.models.loadErrorDesc", { status: response.status }),
+        });
+      }
+    } catch {
+      error({ title: t("admin.models.networkTitle"), description: t("admin.models.saveNetworkDesc") });
+    } finally {
+      setIsSaving(false);
+    }
+  };
+
+  if (isLoading) {
+    return (
+      <div className="flex items-center justify-center py-12">
+        <Loader2 className="size-8 animate-spin text-muted-foreground" />
+      </div>
+    );
+  }
+
+  if (!aiTiers) {
+    return (
+      <Card>
+        <CardContent className="py-8 text-center text-sm text-muted-foreground">
+          {t("admin.models.missingConfig")}
+        </CardContent>
+      </Card>
+    );
+  }
+
+  return (
+    <div className="space-y-6">
+      <div className="flex items-start justify-between gap-4">
+        <div className="flex items-center gap-3">
+          <div className="w-10 h-10 bg-blue-600/20 rounded-lg flex items-center justify-center">
+            <Cpu className="w-5 h-5 text-blue-400" />
+          </div>
+          <div>
+            <h1 className="text-xl font-semibold text-foreground">{t("admin.models.title")}</h1>
+            <p className="text-sm text-muted-foreground">{t("admin.models.subtitle")}</p>
+          </div>
+        </div>
+        <Button onClick={save} disabled={isSaving}>
+          {isSaving ? (
+            <>
+              <Loader2 className="me-2 size-4 animate-spin" /> {t("admin.models.saving")}
+            </>
+          ) : (
+            <>
+              <Save className="me-2 size-4" /> {t("admin.models.save")}
+            </>
+          )}
+        </Button>
+      </div>
+
+      <Card className="overflow-visible">
+        <CardHeader>
+          <CardTitle className="text-base">{t("admin.models.matrixTitle")}</CardTitle>
+          <CardDescription>{t("admin.models.matrixDesc")}</CardDescription>
+        </CardHeader>
+        <CardContent className="space-y-4">
+          {/* Forfait Pro — palier Essentielle */}
+          <div className="rounded-xl border border-border/60 p-4">
+            <div className="mb-3 flex flex-wrap items-center gap-2">
+              <Badge className="bg-accent/10 text-accent border-accent/30">{t("admin.models.proBadge")}</Badge>
+              <Badge variant="secondary">{t("admin.models.essentialBadge")}</Badge>
+              <span className="text-xs text-muted-foreground">{t("admin.models.essentialCost")}</span>
+            </div>
+            <TierEditor
+              tier={aiTiers.essential}
+              onChange={(next) => setAiTiers({ ...aiTiers, essential: next })}
+              catalog={catalog}
+              onLoadCatalog={loadCatalog}
+              catalogLoading={catalogLoading}
+            />
+            <p className="mt-3 text-xs text-muted-foreground">{t("admin.models.sharedTierNote")}</p>
+          </div>
+
+          {/* Forfait Business — palier Premium */}
+          <div className="rounded-xl border border-accent/30 p-4">
+            <div className="mb-3 flex flex-wrap items-center gap-2">
+              <Badge className="bg-blue-500/10 text-blue-600 border-blue-500/30 dark:text-blue-300">
+                {t("admin.models.businessBadge")}
+              </Badge>
+              <Badge variant="secondary">{t("admin.models.premiumBadge")}</Badge>
+              <span className="text-xs text-muted-foreground">{t("admin.models.premiumCost")}</span>
+            </div>
+            <TierEditor
+              tier={aiTiers.premium}
+              onChange={(next) => setAiTiers({ ...aiTiers, premium: next })}
+              catalog={catalog}
+              onLoadCatalog={loadCatalog}
+              catalogLoading={catalogLoading}
+            />
+            <p className="mt-3 text-xs text-muted-foreground">{t("admin.models.premiumReservedNote")}</p>
+          </div>
+
+          {/* Forfait Enterprise — lecture seule */}
+          <div className="rounded-xl border border-border/40 p-4">
+            <div className="mb-1 flex flex-wrap items-center gap-2">
+              <Badge variant="outline">{t("admin.models.enterpriseBadge")}</Badge>
+              <Badge variant="secondary">{t("admin.models.customBadge")}</Badge>
+            </div>
+            <p className="text-xs text-muted-foreground">{t("admin.models.customNote")}</p>
+          </div>
+        </CardContent>
+      </Card>
+    </div>
+  );
+}
--- a/frontend/src/app/admin/settings/page.tsx
+++ b/frontend/src/app/admin/settings/page.tsx
@@ -13,6 +13,7 @@ import { useNotification } from "@/components/ui/notification";
 import { ModelCombobox, ModelOption } from "@/components/ui/model-combobox";
 import { useTranslationStore } from "@/lib/store";
 import { API_BASE } from "@/lib/config";
+import type { AiTiersSettings } from "../types";

 interface ProviderConfig {
  enabled: boolean;
@@ -43,6 +44,8 @@ interface SettingsConfig {
  zai: ProviderConfig;
  mistral: ProviderConfig;
  smtp: SmtpConfig;
+  /** Paliers IA « Modèles & abonnements » — renvoyé tel quel à la sauvegarde. */
+  ai_tiers?: AiTiersSettings;
  fallback_chain: string;
  fallback_chain_classic: string;
  fallback_chain_llm: string;
@@ -70,8 +73,8 @@ const defaultConfig: SettingsConfig = {
  google_cloud: { enabled: false, api_key: "", timeout: 30, max_retries: 3 },
  openai: { enabled: false, api_key: "", timeout: 60, max_retries: 3 },
  ollama: { enabled: false, base_url: "http://localhost:11434", model: "llama3" },
-  openrouter: { enabled: false, api_key: "", model: "deepseek/deepseek-chat" },
-  openrouter_premium: { enabled: false, api_key: "", model: "openai/gpt-4o-mini" },
+  openrouter: { enabled: false, api_key: "", model: "deepseek/deepseek-v4-flash" },
+  openrouter_premium: { enabled: false, api_key: "", model: "anthropic/claude-sonnet-5" },
  zai: { enabled: false, api_key: "", base_url: "https://api.x.ai/v1", model: "grok-2-1212" },
  mistral: { enabled: false, api_key: "", model: "mistral-ocr-latest", timeout: 180 },
  smtp: { enabled: false, host: "", port: 587, username: "", password: "", from_email: "", use_tls: true },
@@ -101,7 +104,6 @@ export default function AdminSettingsPage() {
  const [ollamaModels, setOllamaModels] = useState<OllamaModel[]>([]);
  const [isLoadingModels, setIsLoadingModels] = useState(false);
  const [openaiModels, setOpenaiModels] = useState<ModelOption[]>([]);
-  const [openrouterModels, setOpenrouterModels] = useState<ModelOption[]>([]);
  const [zaiModels, setZaiModels] = useState<ModelOption[]>([]);
  const [loadingModelsProvider, setLoadingModelsProvider] = useState<string | null>(null);
  const [isSendingTestEmail, setIsSendingTestEmail] = useState(false);
@@ -144,6 +146,9 @@ export default function AdminSettingsPage() {
  const saveConfig = async () => {
    setIsSaving(true);
    try {
+      // `config` inclut `ai_tiers` tel que chargé (voir loadConfig) : la
+      // sauvegarde des fournisseurs ne doit jamais réinitialiser les paliers
+      // personnalisés de la page « Modèles & abonnements ».
      const response = await fetch(`${API_BASE}/api/v1/admin/settings`, {
        method: "PUT",
        headers: {
@@ -238,7 +243,6 @@ export default function AdminSettingsPage() {
        const data = await response.json();
        const models: ModelOption[] = data.data || [];
        if (provider === "openai") setOpenaiModels(models);
-        else if (provider === "openrouter") setOpenrouterModels(models);
        else if (provider === "zai") setZaiModels(models);
        info({ title: `${models.length} modèles ${provider} trouvés` });
      } else {
@@ -476,7 +480,7 @@ export default function AdminSettingsPage() {

        <ProviderCard
          title="Traduction IA Essentielle"
-          description="Affichée aux utilisateurs comme 'Traduction IA Essentielle'. Modèles économiques recommandés : deepseek/deepseek-chat, google/gemini-2.0-flash, meta-llama/llama-3.3-70b-instruct. Clé API : openrouter.ai"
+          description="Palier Essentielle (forfait Pro). Le modèle réellement routé se règle dans « Modèles & abonnements ». Gamme : deepseek/deepseek-v4-flash, z-ai/glm-5.3-flash, minimax/minimax-m3. Clé API : openrouter.ai"
          enabled={config.openrouter.enabled}
          onToggle={(enabled) => updateProvider("openrouter", { enabled })}
          onTest={() => testProvider("openrouter")}
@@ -496,24 +500,24 @@ export default function AdminSettingsPage() {
              />
            </div>
            <div className="space-y-2">
-              <Label htmlFor="openrouter-model">Modèle Essentiel</Label>
-              <ModelCombobox
+              <Label htmlFor="openrouter-model">Modèle Essentiel (lecture seule)</Label>
+              <Input
+                id="openrouter-model"
                value={config.openrouter.model || ""}
-                onChange={(v) => updateProvider("openrouter", { model: v })}
-                models={openrouterModels}
-                isLoading={loadingModelsProvider === "openrouter"}
-                onFetchModels={() => fetchModels("openrouter")}
-                providerLabel="OpenRouter"
-                placeholder="deepseek/deepseek-chat"
+                disabled
+                readOnly
+                placeholder="deepseek/deepseek-v4-flash"
              />
-              <p className="text-xs text-muted-foreground">Recommandé : <code>deepseek/deepseek-chat</code> (~€0.04/doc)</p>
+              <p className="text-xs text-amber-500">
+                Piloté par la page <strong>Modèles &amp; abonnements</strong> (matrice par forfait).
+              </p>
            </div>
          </div>
        </ProviderCard>

        <ProviderCard
          title="Traduction IA Premium"
-          description="Affichée aux utilisateurs comme 'Traduction IA Premium'. Modèles haute qualité : openai/gpt-4o, anthropic/claude-sonnet-4.6, google/gemini-3.5-pro. Partage la même clé OpenRouter."
+          description="Palier Premium (forfait Business). Le modèle réellement routé se règle dans « Modèles & abonnements ». Gamme : anthropic/claude-sonnet-5, deepseek/deepseek-v4-pro, z-ai/glm-5.3. Partage la même clé OpenRouter."
          enabled={config.openrouter_premium.enabled}
          onToggle={(enabled) => updateProvider("openrouter_premium", { enabled })}
          onTest={() => testProvider("openrouter_premium")}
@@ -521,21 +525,19 @@ export default function AdminSettingsPage() {
          testMessage={testMessages.openrouter_premium}
          envKeySet={envInfo.openrouter_premium}
        >
-          <div className="space-y-2">
-            <Label htmlFor="openrouter-premium-model">Modèle Premium</Label>
-            <ModelCombobox
-              value={config.openrouter_premium.model || ""}
-              onChange={(v) => updateProvider("openrouter_premium", { model: v })}
-              models={openrouterModels}
-              isLoading={loadingModelsProvider === "openrouter"}
-              onFetchModels={() => fetchModels("openrouter")}
-              providerLabel="OpenRouter"
-              placeholder="anthropic/claude-sonnet-4.6"
-            />
-            <p className="text-xs text-muted-foreground">
-              Recommandé : <code>anthropic/claude-sonnet-4.6</code> (~€0.20/doc) ou <code>openai/gpt-4o</code> (~€0.30/doc)
-            </p>
-          </div>
+            <div className="space-y-2">
+              <Label htmlFor="openrouter-premium-model">Modèle Premium (lecture seule)</Label>
+              <Input
+                id="openrouter-premium-model"
+                value={config.openrouter_premium.model || ""}
+                disabled
+                readOnly
+                placeholder="anthropic/claude-sonnet-5"
+              />
+              <p className="text-xs text-amber-500">
+                Piloté par la page <strong>Modèles &amp; abonnements</strong> (matrice par forfait).
+              </p>
+            </div>
        </ProviderCard>

        <ProviderCard
--- a/frontend/src/app/admin/stats/page.tsx
+++ b/frontend/src/app/admin/stats/page.tsx
@@ -4,6 +4,7 @@ import { useState } from "react";
 import { BarChart3, RefreshCw, Loader2, AlertCircle, Info } from "lucide-react";
 import { Button } from "@/components/ui/button";
 import { StatsOverview } from "../StatsOverview";
+import { RevenueOverview } from "../RevenueOverview";
 import { TopUsersTable } from "../TopUsersTable";
 import { ProviderBreakdownChart } from "../ProviderBreakdownChart";
 import { FormatBreakdownChart } from "../FormatBreakdownChart";
@@ -33,10 +34,10 @@ export default function StatsPage() {
            </div>
            <div>
              <h1 className="text-xl font-semibold text-foreground">
-                Statistiques de Traduction
+                Statistiques
              </h1>
              <p className="text-sm text-muted-foreground">
-                Analyse des traductions et patterns d&apos;utilisation
+                Revenus, MRR, traductions et patterns d&apos;utilisation
              </p>
            </div>
          </div>
@@ -74,6 +75,8 @@ export default function StatsPage() {
          </div>
        )}

+        <RevenueOverview />
+
        <StatsOverview data={data} isLoading={isLoading} />

        <div className="grid gap-6 md:grid-cols-2">
--- a/frontend/src/app/admin/types.ts
+++ b/frontend/src/app/admin/types.ts
@@ -78,3 +78,90 @@ export interface TranslationStatsResponse {
    generated_at: string;
  };
 }
+
+/* ── Modèles & abonnements (paliers IA) ─────────────────────── */
+
+export interface AiTierSettings {
+  /** Modèles actifs — l'ordre est la priorité de secours. */
+  models: string[];
+  /** Modèle par défaut ; vide = premier de la liste = défaut du plan. */
+  default_model: string;
+}
+
+export interface AiTiersSettings {
+  essential: AiTierSettings;
+  premium: AiTierSettings;
+}
+
+/* ── Statistiques enrichies (revenus, MRR, liste d'attente) ── */
+
+export interface AdminStatsResponse {
+  users: {
+    total: number;
+    active_this_month: number;
+    by_plan: Record<string, number>;
+  };
+  translations: {
+    docs_this_month: number;
+    pages_this_month: number;
+  };
+  revenue: {
+    collected_total: number;
+    collected_30d: number;
+    credits_purchased: number;
+    payments_30d: number;
+    currency: string;
+  };
+  mrr_estimated: {
+    total: number;
+    by_plan: Record<string, number>;
+  };
+  ai_tier_usage: {
+    essential: number;
+    premium: number;
+    classic: number;
+    other: number;
+  };
+  waitlist_count: number;
+  cache: Record<string, unknown>;
+  config: Record<string, unknown>;
+}
+
+/* ── Marketing (relances email) ─────────────────────────────── */
+
+export interface MarketingAudiences {
+  audiences: Record<string, number>;
+  unsubscribed_count: number;
+}
+
+export interface MarketingAudiencesResponse {
+  data: MarketingAudiences;
+  meta: Record<string, unknown>;
+}
+
+export interface MarketingEmailHistoryEntry {
+  id: string;
+  sent_at: string;
+  completed_at?: string;
+  audience: string;
+  subject: string;
+  test_mode: boolean;
+  test_recipient?: string;
+  recipients_total: number;
+  sent_count: number;
+  failed_count: number;
+  failed: { email: string; error: string }[];
+  content_hash?: string;
+}
+
+export interface MarketingEmailSendResponse {
+  data: {
+    test_mode: boolean;
+    sent?: number;
+    failed?: number;
+    recipient?: string;
+    queued?: number;
+    audience?: string;
+  };
+  meta: Record<string, unknown>;
+}
--- a/frontend/src/app/pricing/page.tsx
+++ b/frontend/src/app/pricing/page.tsx
@@ -922,7 +922,7 @@ export default function PricingPage() {
              </div>
              <div className="flex flex-wrap gap-2 text-xs">
                <span className="px-2 py-1 bg-muted rounded">{t('pricing.aiModels.essential.context')}</span>
-                <span className="px-2 py-1 bg-muted rounded">$0.25/$0.38 per 1M</span>
+                <span className="px-2 py-1 bg-muted rounded">{t('pricing.aiModels.essential.price')}</span>
                <span className="px-2 py-1 bg-emerald-500/10 text-emerald-600 dark:text-emerald-400 rounded">{t('pricing.aiModels.essential.value')}</span>
              </div>
            </div>
@@ -933,11 +933,11 @@ export default function PricingPage() {
                <Badge className="ml-auto text-xs bg-accent/10 text-accent border-accent/30">{t('pricing.aiModels.premium.plan')}</Badge>
              </div>
              <div className="text-sm text-muted-foreground mb-3">
-                {t('pricing.aiModels.premium.descPrefix')} <strong className="text-foreground">Claude Sonnet 4.6</strong> {t('pricing.aiModels.premium.descSuffix')}
+                {t('pricing.aiModels.premium.descPrefix')} <strong className="text-foreground">{t('pricing.aiModels.premium.modelName')}</strong> {t('pricing.aiModels.premium.descSuffix')}
              </div>
              <div className="flex flex-wrap gap-2 text-xs">
                <span className="px-2 py-1 bg-muted rounded">{t('pricing.aiModels.premium.context')}</span>
-                <span className="px-2 py-1 bg-muted rounded">$3.00/$15.00 per 1M</span>
+                <span className="px-2 py-1 bg-muted rounded">{t('pricing.aiModels.premium.alternatives')}</span>
                <span className="px-2 py-1 bg-accent/10 text-accent rounded">{t('pricing.aiModels.premium.precision')}</span>
              </div>
            </div>
--- a/frontend/src/lib/i18n/messages/ar/admin.json
+++ b/frontend/src/lib/i18n/messages/ar/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "المزوّدون",
  "admin.nav.system": "النظام",
  "admin.nav.logs": "السجلات",
+  "admin.nav.stats": "الإحصائيات",
  "admin.users.title": "إدارة المستخدمين",
  "admin.users.subtitle": "عرض حسابات المستخدمين وإدارتها",
  "admin.users.planUpdated": "تم تحديث الخطة",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "في انتظار البيانات...",
  "admin.system.purging": "جارٍ الحذف...",
  "admin.system.clean": "تنظيف",
-  "admin.system.purge": "حذف"
+  "admin.system.purge": "حذف",
+  "admin.nav.models": "النماذج والاشتراكات",
+  "admin.nav.marketing": "التسويق",
+  "admin.marketing.audience": "الجمهور",
+  "admin.marketing.audienceAllUsers": "جميع الحسابات",
+  "admin.marketing.audienceInactive": "حسابات غير نشطة منذ 30 يومًا",
+  "admin.marketing.audiencePlaceholder": "اختر جمهورًا",
+  "admin.marketing.audiencePlanBusiness": "خطة Business",
+  "admin.marketing.audiencePlanEnterprise": "خطة Enterprise",
+  "admin.marketing.audiencePlanFree": "الخطة المجانية",
+  "admin.marketing.audiencePlanPro": "خطة Pro",
+  "admin.marketing.audiencePlanStarter": "خطة Starter",
+  "admin.marketing.audienceWaitlist": "قائمة الانتظار",
+  "admin.marketing.badgeReal": "فعلي",
+  "admin.marketing.badgeTest": "اختباري",
+  "admin.marketing.confirmSend": "إرسال هذه الرسالة إلى {count} مستلم («{audience}»)؟",
+  "admin.marketing.footerNote": "يُضاف تلقائيًا تذييل يحتوي على رابط إلغاء الاشتراك عند الإرسال.",
+  "admin.marketing.hidePreview": "إخفاء المعاينة",
+  "admin.marketing.history": "سجل الإرسالات",
+  "admin.marketing.historyFailedSuffix": "، {count} فاشلة",
+  "admin.marketing.historySent": "{sent}/{total} مُرسلة",
+  "admin.marketing.html": "محتوى HTML",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "أدخل الموضوع وHTML.",
+  "admin.marketing.incompleteTitle": "محتوى غير مكتمل",
+  "admin.marketing.loadErrorTitle": "خطأ في التحميل",
+  "admin.marketing.localCheckUnavailable": "التحقق المحلي غير متاح في هذا المتصفح: سيتحقق الخادم من الإرسال الاختباري المسبق.",
+  "admin.marketing.networkDesc": "لا يمكن الوصول إلى الخادم الخلفي.",
+  "admin.marketing.newCampaign": "حملة جديدة",
+  "admin.marketing.noHistory": "لا توجد إرسالات مسجلة.",
+  "admin.marketing.preview": "معاينة",
+  "admin.marketing.previewEmpty": "معاينة: الصق HTML بجانب هذه اللوحة.",
+  "admin.marketing.previewTitle": "معاينة البريد",
+  "admin.marketing.recipientsAvailable": "{count} مستلم متاح بعد استثناء الملغين لاشتراكهم.",
+  "admin.marketing.sendReal": "الإرسال إلى {count} مستلم",
+  "admin.marketing.sendRefusedTitle": "تم رفض الإرسال",
+  "admin.marketing.sendTest": "إرسال اختباري",
+  "admin.marketing.sendingDesc": "{count} رسالة في قائمة الإرسال (فاصل 0,2 ثانية). الملغون لاشتراكهم مستثنون.",
+  "admin.marketing.sendingTitle": "الإرسال جارٍ",
+  "admin.marketing.subject": "الموضوع",
+  "admin.marketing.subjectPlaceholder": "مثال: ترجمتك بانتظارك — خصم 20٪ لمدة 7 أيام",
+  "admin.marketing.subtitle": "إرسال رسائل إلى جمهور محدد، مع إرسال اختباري إلزامي ورابط إلغاء اشتراك تلقائي وسجل كامل.",
+  "admin.marketing.testDoneDesc": "وصل البريد إلى {email}",
+  "admin.marketing.testDoneTitle": "تم الإرسال الاختباري",
+  "admin.marketing.testEmail": "بريد اختباري (اختياري)",
+  "admin.marketing.testEmailPlaceholder": "وإلا: عنوان المرسل SMTP",
+  "admin.marketing.testFailedTitle": "فشل الإرسال الاختباري",
+  "admin.marketing.testRequiredHint": "يتطلب الإرسال الفعلي إرسالًا اختباريًا مسبقًا لنفس هذا المحتوى (يتحقق منه الخادم).",
+  "admin.marketing.testValidated": "تم التحقق من الإرسال الاختباري لهذا المحتوى.",
+  "admin.marketing.title": "التسويق — حملات البريد الإلكتروني",
+  "admin.marketing.unsubscribedNote": "سيتم استثناء {count} من الملغين لاشتراكهم من كل إرسال.",
+  "admin.models.addModel": "إضافة نموذج",
+  "admin.models.businessBadge": "خطة Business",
+  "admin.models.cancel": "إلغاء",
+  "admin.models.catalog": "كتالوج OpenRouter",
+  "admin.models.catalogErrorDesc": "تعذر تحميل كتالوج OpenRouter.",
+  "admin.models.catalogErrorTitle": "الكتالوج غير متاح",
+  "admin.models.customBadge": "مستوى مخصص",
+  "admin.models.customNote": "نماذج حسب الطلب، تُحدد حالة بحالة مع العميل.",
+  "admin.models.defaultBadge": "افتراضي",
+  "admin.models.emptyTier": "لا نماذج: ستُستخدم المجموعة الرسمية للخطة.",
+  "admin.models.enterpriseBadge": "خطة Enterprise",
+  "admin.models.essentialBadge": "مستوى الذكاء الاصطناعي الأساسي",
+  "admin.models.essentialCost": "— يُحتسب بمعامل تكلفة 1",
+  "admin.models.fallbackFirst": "(احتياط رقم 1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — تحقق من رمز المسؤول.",
+  "admin.models.loadErrorTitle": "خطأ في التحميل",
+  "admin.models.matrixDesc": "النماذج النشطة لكل مستوى — ترتيب القائمة هو أولوية الاحتياط. يختار زر الراديو النموذج الافتراضي، ويُطبق فورًا دون إعادة نشر.",
+  "admin.models.matrixTitle": "مصفوفة النماذج حسب الخطة",
+  "admin.models.missingConfig": "لم يتم العثور على إعدادات مستويات الذكاء الاصطناعي. أعد تحميل الصفحة.",
+  "admin.models.moveDown": "تحريك لأسفل",
+  "admin.models.moveUp": "تحريك لأعلى",
+  "admin.models.networkDesc": "لا يمكن الوصول إلى الخادم الخلفي.",
+  "admin.models.networkTitle": "خطأ في الشبكة",
+  "admin.models.premiumBadge": "مستوى الذكاء الاصطناعي المتميز",
+  "admin.models.premiumCost": "— يُحتسب بمعامل تكلفة 5",
+  "admin.models.premiumReservedNote": "محصور في خطتي Business وEnterprise (محرك «openrouter_premium»).",
+  "admin.models.proBadge": "خطة Pro",
+  "admin.models.remove": "إزالة",
+  "admin.models.save": "حفظ",
+  "admin.models.saveErrorTitle": "خطأ في الحفظ",
+  "admin.models.saveNetworkDesc": "تعذر حفظ الإعدادات.",
+  "admin.models.savedDesc": "يُستخدم النموذج الافتراضي الجديد بدءًا من الترجمة التالية، دون إعادة نشر.",
+  "admin.models.savedTitle": "تم حفظ النماذج",
+  "admin.models.saving": "جارٍ الحفظ...",
+  "admin.models.setDefault": "تعيين {model} كنموذج افتراضي",
+  "admin.models.sharedTierNote": "مستوى مشترك: تستخدم خطة Business المستوى الأساسي أيضًا لمحركها «openrouter» — القائمة أعلاه تنطبق على الخطتين.",
+  "admin.models.subtitle": "مصفوفة الخطط → مستويات الذكاء الاصطناعي. النموذج المستخدم فعليًا في كل ترجمة هو نموذج مستوى الخطة: خطة Pro لا يمكنها أبدًا تشغيل نموذج Premium.",
+  "admin.models.title": "النماذج والاشتراكات",
+  "admin.stats.aiTiers": "مستويات الذكاء الاصطناعي (30 يومًا)",
+  "admin.stats.includingCredits": "منها {amount} من الرصيد الإضافي",
+  "admin.stats.mrr": "الإيراد الشهري المتكرر المقدر",
+  "admin.stats.noTranslationsYet": "لا توجد ترجمات مسجلة خلال آخر 30 يومًا: ستكتمل توزيعات المستويات مع الترجمات القادمة.",
+  "admin.stats.payments30": "{count} دفعة خلال 30 يومًا",
+  "admin.stats.refreshing": "جارٍ التحديث...",
+  "admin.stats.revenue30": "الإيرادات (30 يومًا)",
+  "admin.stats.revenueTotal": "إجمالي الإيرادات المحصلة",
+  "admin.stats.tiersSub": "أساسي / متميز · كلاسيكي {classic} · آخر {other}",
+  "admin.stats.unavailable": "إحصائيات الأعمال غير متاحة ({error}).",
+  "admin.stats.waitlist": "قائمة الانتظار",
+  "admin.stats.waitlistSub": "مُدرجون في قائمة الانتظار"
 }
--- a/frontend/src/lib/i18n/messages/ar/landing.json
+++ b/frontend/src/lib/i18n/messages/ar/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "للمحترفين المتميزين",
  "landing.pricing.pro.f1": "200 مستند / شهر",
  "landing.pricing.pro.f2": "حتى 200 صفحة لكل مستند",
-  "landing.pricing.pro.f3": "ترجمة بالذكاء الاصطناعي",
+  "landing.pricing.pro.f3": "ذكاء اصطناعي أساسي: DeepSeek V4 Flash، GLM-5.3 Flash، MiniMax M3",
  "landing.pricing.pro.f4": "Google مشمولان",
  "landing.pricing.pro.f5": "قواميس وأوامر مخصصة",
  "landing.pricing.pro.f6": "دعم ذو أولوية",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "للفرق ذات الاحتياجات الكبيرة",
  "landing.pricing.business.f1": "1,000 مستند / شهر",
  "landing.pricing.business.f2": "حتى 500 صفحة لكل مستند",
-  "landing.pricing.business.f3": "ذكاء اصطناعي متميز (Claude)",
+  "landing.pricing.business.f3": "ذكاء اصطناعي متميز (Claude 5)",
  "landing.pricing.business.f4": "جميع المزودين + واجهة برمجة التطبيقات",
  "landing.pricing.business.f5": "خطاطف الويب والأتمتة",
  "landing.pricing.business.f6": "5 مقاعد للفريق",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "ملفات DOCX, XLSX, PPTX أو PDF مدعومة",
  "landing.translate.aiAnalysis": "تحليل AI نشط",
  "landing.translate.processing": "جاري المعالجة",
-  "landing.translate.preservingLayout": "جاري الحفاظ على التنسيق"
+  "landing.translate.preservingLayout": "جاري الحفاظ على التنسيق",
+  "landing.beforeAfter.seal": "نفس التخطيط، كلمة بكلمة",
+  "landing.beforeAfter.proof1": "مخططات SmartArt يُعاد بناؤها، وليس تسطيحها",
+  "landing.beforeAfter.proof2": "سلاسل الرسوم البيانية ومحاورها مترجمة",
+  "landing.beforeAfter.proof3": "جداول المحتويات يُعاد إنشاؤها باللغة الهدف",
+  "landing.beforeAfter.targetTitle": "المواصفات الفنية — معالجة الهواء",
+  "landing.beforeAfter.targetPara1": "يجب تشغيل نظام التهوية قبل نهاية الربع الثاني.",
+  "landing.beforeAfter.targetPara2Before": "",
+  "landing.beforeAfter.targetTerm": "وحدة معالجة الهواء",
+  "landing.beforeAfter.targetPara2After": "يجب أن تتوافق مع فئة الترشيح F7 وفقًا للجدول.",
+  "landing.beforeAfter.targetItem1": "الحمل الحراري: 42 كيلوواط عند التدفق الاسمي",
+  "landing.beforeAfter.targetItem2": "مستوى الصوت أقل من 45 dB(A) على مسافة 3 أمتار",
+  "landing.formats.pill": "التوافق",
+  "landing.hero.visualCaption": "نفس التخطيط، لغة جديدة — لا شيء آخر يتغير",
+  "landing.pricing.free.name": "مجاني",
+  "landing.pricing.free.desc": "مثالي لاكتشاف التطبيق",
+  "landing.pricing.free.cta": "اختر هذه الخطة",
+  "landing.pricing.enterprise.name": "المؤسسات",
+  "landing.pricing.enterprise.desc": "حلول مخصصة للمؤسسات الكبيرة",
+  "landing.pricing.enterprise.cta": "اتصل بنا",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/ar/pricing.json
+++ b/frontend/src/lib/i18n/messages/ar/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "سجل 90 يومًا",
  "pricing.plans.business.feat1": "1 000 مستند / شهر",
  "pricing.plans.business.feat2": "حتى 500 صفحة لكل مستند",
-  "pricing.plans.business.feat3": "AI أساسية + متميزة (Claude Haiku)",
+  "pricing.plans.business.feat3": "الذكاء الاصطناعي الأساسي + المتميز (Claude 5)",
  "pricing.plans.business.feat4": "جميع مزوّدي الترجمة",
  "pricing.plans.business.feat5": "ملفات حتى 50 ميغابايت",
  "pricing.plans.business.feat6": "وصول API (10 000 استدعاء/شهر)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "سجل سنة واحدة",
  "pricing.plans.business.feat10": "تحليلات متقدمة",
  "pricing.plans.enterprise.feat1": "مستندات بلا حدود",
-  "pricing.plans.enterprise.feat2": "جميع نماذج AI (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "جميع نماذج AI (Claude 5، DeepSeek V4 Pro، GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "نشر محلي أو سحابة مخصصة",
  "pricing.plans.enterprise.feat4": "SLA 99.9% مضمون",
  "pricing.plans.enterprise.feat5": "دعم مخصص على مدار الساعة",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "ذكاء اصطناعي متعدد المسارات فائق السرعة",
  "pricing.trust.availability.title": "متاح على مدار الساعة",
  "pricing.trust.availability.sub": "ضمان وقت تشغيل 99.9%",
-  "pricing.aiModels.title": "نماذج الذكاء الاصطناعي — مارس 2026",
+  "pricing.aiModels.title": "نماذج الذكاء الاصطناعي — سبتمبر 2026",
  "pricing.aiModels.essential.title": "ترجمة AI أساسية",
  "pricing.aiModels.essential.plan": "خطة Pro",
  "pricing.aiModels.essential.descPrefix": "مبني على",
-  "pricing.aiModels.essential.descSuffix": "— أنموذج الذكاء الاصطناعي الأكثر فعالية من حيث التكلفة لعام 2026. جودة مماثلة للنماذج المتقدمة بتكلفة أقل بكثير.",
-  "pricing.aiModels.essential.modelName": "نموذج الذكاء الاصطناعي الأساسي",
-  "pricing.aiModels.essential.context": "163K رمز سياقي",
-  "pricing.aiModels.essential.value": "قيمة ممتازة مقابل المال",
+  "pricing.aiModels.essential.descSuffix": "— نماذج الذكاء الاصطناعي الأفضل قيمة مقابل المال لعام 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash وGLM-5.3 Flash وMiniMax M3",
+  "pricing.aiModels.essential.context": "سياق يصل إلى 1.3 مليون رمز (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "أفضل قيمة مقابل السعر",
  "pricing.aiModels.premium.title": "ترجمة AI متميزة",
  "pricing.aiModels.premium.plan": "خطة Business",
  "pricing.aiModels.premium.descPrefix": "مبني على",
  "pricing.aiModels.premium.descSuffix": "من Anthropic — دقة عالية في المستندات القانونية والطبية والتقنية المعقدة.",
-  "pricing.aiModels.premium.context": "200K رمز سياقي",
+  "pricing.aiModels.premium.context": "مليون رمز سياقي",
  "pricing.aiModels.premium.precision": "أعلى دقة",
  "pricing.faq.title": "الأسئلة الشائعة",
  "pricing.faq.q1": "هل يمكنني تغيير خطتي في أي وقت؟",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "ما هي «ترجمة AI الأساسية»؟",
  "pricing.faq.a2": "إنه محرك الذكاء الاصطناعي الخاص بنا. يفهم سياق مستنداتك، ويحافظ على التخطيط، ويتعامل مع المصطلحات التقنية بشكل أفضل بكثير من الترجمة الكلاسيكية.",
  "pricing.faq.q3": "ما الفرق بين AI الأساسية وAI المتميزة؟",
-  "pricing.faq.a3": "الذكاء الاصطناعي الأساسي يستخدم نموذجًا محسّنًا (قيمة ممتازة مقابل المال). الذكاء الاصطناعي المتميز يستخدم Claude 3.5 Haiku من Anthropic، أكثر دقة في المستندات القانونية والطبية والتقنية المعقدة.",
+  "pricing.faq.a3": "الذكاء الاصطناعي الأساسي يعمل بنماذج DeepSeek V4 Flash وGLM-5.3 Flash وMiniMax M3 (قيمة ممتازة مقابل المال). الذكاء الاصطناعي المتميز يستخدم Claude 5 من Anthropic، أكثر دقة في المستندات القانونية والطبية والتقنية المعقدة.",
  "pricing.faq.q4": "هل تُحفظ مستنداتي بعد الترجمة؟",
  "pricing.faq.a4": "الملفات المترجمة متاحة حسب خطتك (30 يومًا لـ Starter، 90 يومًا لـ Pro، سنة واحدة لـ Business). وهي مشفرة أثناء التخزين والنقل.",
  "pricing.faq.q5": "ماذا يحدث إذا تجاوزت حصتي الشهرية؟",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "خطأ أثناء إنشاء الدفع.",
  "pricing.dashboard": "لوحة التحكم",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "ابتداءً من 0.09 دولار لكل مليون رمز",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "بدائل: DeepSeek V4 Pro وGLM-5.3",
+  "pricing.confirm.title": "أكّد اشتراكك",
+  "pricing.confirm.subtitle": "سيتم تحويلك إلى مزود الدفع الآمن لدينا.",
+  "pricing.confirm.cancel": "إلغاء",
+  "pricing.confirm.cta": "المتابعة إلى الدفع",
+  "pricing.confirm.year": "سنة",
+  "pricing.confirm.month": "شهر",
+  "pricing.confirm.monthlyEquivalent": "فوترة سنوية — أي {price} € / شهر.",
+  "pricing.confirm.secureNote": "يمكنك الإلغاء في أي وقت من ملفك الشخصي. تتم معالجة الدفع عبر Stripe؛ ولا تمر بيانات بطاقتك أبدًا عبر خوادمنا.",
+  "pricing.enterprise.subject": "عرض المؤسسات",
+  "pricing.enterpriseBand.cta": "اتصل بنا",
+  "pricing.enterpriseBand.text": "حجم كبير، محركات مخصصة، خيارات تثبيت داخلي — لنتحدث.",
+  "pricing.error.server": "خطأ في الخادم {status}",
+  "pricing.freeBand.cta": "ابدأ مجانًا",
+  "pricing.freeBand.text": "تريد التجربة فقط؟ ابدأ مجانًا — 5 مستندات شهريًا، بدون بطاقة.",
+  "pricing.header.titleBase": "خطة لكل",
+  "pricing.header.titleAccent": "احتياج",
+  "pricing.toast.close": "إغلاق",
+  "pricing.aiModels.premium.price": "2 $ / 10 $ لكل مليون رمز"
 }
--- a/frontend/src/lib/i18n/messages/ar/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/ar/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "وصول موحد إلى أفضل النماذج مفتوحة المصدر المحسّنة للترجمة.",
  "providerTheme.openrouter_premium.badge": "فائق",
  "providerTheme.openrouter_premium.subBadge": "السياق الأقصى",
-  "providerTheme.openrouter_premium.desc": "مدعوم من أحدث النماذج (GPT-4o، Claude Sonnet 4.6) للمستندات الطويلة.",
+  "providerTheme.openrouter_premium.desc": "مدعوم من أحدث النماذج (GPT-4o، Claude 5) للمستندات الطويلة.",
  "providerTheme.zai.badge": "متخصص",
  "providerTheme.zai.subBadge": "المال والقانون",
  "providerTheme.zai.desc": "نموذج مضبوط بدقة للمصطلحات التجارية المتطلبة (قانونية، مالية).",
--- a/frontend/src/lib/i18n/messages/de/admin.json
+++ b/frontend/src/lib/i18n/messages/de/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Anbieter",
  "admin.nav.system": "System",
  "admin.nav.logs": "Protokolle",
+  "admin.nav.stats": "Statistiken",
  "admin.users.title": "Benutzerverwaltung",
  "admin.users.subtitle": "Benutzerkonten anzeigen und verwalten",
  "admin.users.planUpdated": "Plan aktualisiert",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "Warte auf Daten...",
  "admin.system.purging": "Bereinigung läuft...",
  "admin.system.clean": "Bereinigen",
-  "admin.system.purge": "Bereinigen"
+  "admin.system.purge": "Bereinigen",
+  "admin.nav.models": "Modelle & Abonnements",
+  "admin.nav.marketing": "Marketing",
+  "admin.marketing.audience": "Zielgruppe",
+  "admin.marketing.audienceAllUsers": "Alle Konten",
+  "admin.marketing.audienceInactive": "Seit 30 Tagen inaktive Konten",
+  "admin.marketing.audiencePlaceholder": "Zielgruppe wählen",
+  "admin.marketing.audiencePlanBusiness": "Business-Tarif",
+  "admin.marketing.audiencePlanEnterprise": "Enterprise-Tarif",
+  "admin.marketing.audiencePlanFree": "Kostenlos-Tarif",
+  "admin.marketing.audiencePlanPro": "Pro-Tarif",
+  "admin.marketing.audiencePlanStarter": "Starter-Tarif",
+  "admin.marketing.audienceWaitlist": "Warteliste",
+  "admin.marketing.badgeReal": "echt",
+  "admin.marketing.badgeTest": "Test",
+  "admin.marketing.confirmSend": "Diese E-Mail an {count} Empfänger senden („{audience}“)?",
+  "admin.marketing.footerNote": "Eine Fußzeile mit dem Abmeldelink wird beim Senden automatisch ergänzt.",
+  "admin.marketing.hidePreview": "Vorschau ausblenden",
+  "admin.marketing.history": "Sendungsverlauf",
+  "admin.marketing.historyFailedSuffix": ", {count} Fehler",
+  "admin.marketing.historySent": "{sent}/{total} gesendet",
+  "admin.marketing.html": "HTML-Inhalt",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Bitte Betreff und HTML ausfüllen.",
+  "admin.marketing.incompleteTitle": "Unvollständiger Inhalt",
+  "admin.marketing.loadErrorTitle": "Ladefehler",
+  "admin.marketing.localCheckUnavailable": "Lokale Prüfung in diesem Browser nicht möglich: Der Server kontrolliert die vorherige Testsendung.",
+  "admin.marketing.networkDesc": "Backend nicht erreichbar.",
+  "admin.marketing.newCampaign": "Neue Kampagne",
+  "admin.marketing.noHistory": "Keine Sendung erfasst.",
+  "admin.marketing.preview": "Vorschau",
+  "admin.marketing.previewEmpty": "Vorschau: HTML daneben einfügen.",
+  "admin.marketing.previewTitle": "E-Mail-Vorschau",
+  "admin.marketing.recipientsAvailable": "{count} Empfänger nach Ausschluss der Abmeldungen verfügbar.",
+  "admin.marketing.sendReal": "An {count} Empfänger senden",
+  "admin.marketing.sendRefusedTitle": "Sendung abgelehnt",
+  "admin.marketing.sendTest": "Testsendung",
+  "admin.marketing.sendingDesc": "{count} E-Mail(s) in der Warteschlange (0,2 s Abstand). Abmeldungen werden ausgeschlossen.",
+  "admin.marketing.sendingTitle": "Sendung läuft",
+  "admin.marketing.subject": "Betreff",
+  "admin.marketing.subjectPlaceholder": "z. B.: Ihre Übersetzung wartet — 20 % für 7 Tage",
+  "admin.marketing.subtitle": "E-Mails an eine Zielgruppe senden, mit Pflicht-Testsendung, automatischem Abmeldelink und vollständigem Verlauf.",
+  "admin.marketing.testDoneDesc": "E-Mail erhalten an {email}",
+  "admin.marketing.testDoneTitle": "Testsendung erfolgt",
+  "admin.marketing.testEmail": "Test-E-Mail (optional)",
+  "admin.marketing.testEmailPlaceholder": "Sonst: SMTP-Absenderadresse",
+  "admin.marketing.testFailedTitle": "Testsendung fehlgeschlagen",
+  "admin.marketing.testRequiredHint": "Eine echte Sendung erfordert eine vorherige Testsendung exakt dieses Inhalts (vom Server geprüft).",
+  "admin.marketing.testValidated": "Testsendung für diesen Inhalt bestätigt.",
+  "admin.marketing.title": "Marketing — E-Mail-Kampagnen",
+  "admin.marketing.unsubscribedNote": "{count} abgemeldete Kontakte werden von jeder Sendung ausgeschlossen.",
+  "admin.models.addModel": "Modell hinzufügen",
+  "admin.models.businessBadge": "Business-Tarif",
+  "admin.models.cancel": "Abbrechen",
+  "admin.models.catalog": "OpenRouter-Katalog",
+  "admin.models.catalogErrorDesc": "OpenRouter-Katalog konnte nicht geladen werden.",
+  "admin.models.catalogErrorTitle": "Katalog nicht verfügbar",
+  "admin.models.customBadge": "Individuelle Stufe",
+  "admin.models.customNote": "Sondermodelle, im Einzelfall mit dem Kunden vereinbart.",
+  "admin.models.defaultBadge": "Standard",
+  "admin.models.emptyTier": "Kein Modell: Die offizielle Modellrange des Tarifs wird verwendet.",
+  "admin.models.enterpriseBadge": "Enterprise-Tarif",
+  "admin.models.essentialBadge": "KI-Stufe Basis",
+  "admin.models.essentialCost": "— Abrechnungsfaktor 1",
+  "admin.models.fallbackFirst": "(Fallback Nr. 1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — prüfen Sie Ihr Admin-Token.",
+  "admin.models.loadErrorTitle": "Ladefehler",
+  "admin.models.matrixDesc": "Aktive Modelle pro Stufe — die Reihenfolge der Liste ist die Fallback-Priorität. Der Radioknopf wählt das Standardmodell, sofort und ohne neues Deployment wirksam.",
+  "admin.models.matrixTitle": "Modellmatrix pro Tarif",
+  "admin.models.missingConfig": "KI-Stufen-Konfiguration nicht gefunden. Bitte Seite neu laden.",
+  "admin.models.moveDown": "Nach unten",
+  "admin.models.moveUp": "Nach oben",
+  "admin.models.networkDesc": "Backend nicht erreichbar.",
+  "admin.models.networkTitle": "Netzwerkfehler",
+  "admin.models.premiumBadge": "KI-Stufe Premium",
+  "admin.models.premiumCost": "— Abrechnungsfaktor 5",
+  "admin.models.premiumReservedNote": "Den Tarifen Business und Enterprise vorbehalten (Engine „openrouter_premium“).",
+  "admin.models.proBadge": "Pro-Tarif",
+  "admin.models.remove": "Entfernen",
+  "admin.models.save": "Speichern",
+  "admin.models.saveErrorTitle": "Speicherfehler",
+  "admin.models.saveNetworkDesc": "Konfiguration konnte nicht gespeichert werden.",
+  "admin.models.savedDesc": "Das neue Standardmodell wird ab der nächsten Übersetzung verwendet — ohne neues Deployment.",
+  "admin.models.savedTitle": "Modelle gespeichert",
+  "admin.models.saving": "Speichern...",
+  "admin.models.setDefault": "{model} als Standardmodell festlegen",
+  "admin.models.sharedTierNote": "Gemeinsame Stufe: Der Business-Tarif nutzt die Basis-Stufe ebenfalls für seine Engine „openrouter“ — die obige Liste gilt für beide Tarife.",
+  "admin.models.subtitle": "Matrix Tarife → KI-Stufen. Das tatsächlich pro Übersetzung verwendete Modell ist das der KI-Stufe des Tarifs: Ein Pro-Tarif kann nie ein Premium-Modell auslösen.",
+  "admin.models.title": "Modelle & Abonnements",
+  "admin.stats.aiTiers": "KI-Stufen (30 T)",
+  "admin.stats.includingCredits": "davon {amount} Credits",
+  "admin.stats.mrr": "Geschätzter MRR",
+  "admin.stats.noTranslationsYet": "Keine Übersetzungen der letzten 30 Tage in der Datenbank: Die Stufenverteilung füllt sich mit den nächsten Übersetzungen.",
+  "admin.stats.payments30": "{count} Zahlung(en) in 30 Tagen",
+  "admin.stats.refreshing": "Aktualisierung...",
+  "admin.stats.revenue30": "Umsatz (30 Tage)",
+  "admin.stats.revenueTotal": "Eingenommener Umsatz (gesamt)",
+  "admin.stats.tiersSub": "Basis / Premium · klassisch {classic} · sonstige {other}",
+  "admin.stats.unavailable": "Geschäftszahlen nicht verfügbar ({error}).",
+  "admin.stats.waitlist": "Warteliste",
+  "admin.stats.waitlistSub": "Einträge auf der Warteliste"
 }
--- a/frontend/src/lib/i18n/messages/de/landing.json
+++ b/frontend/src/lib/i18n/messages/de/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "Für anspruchsvolle Profis",
  "landing.pricing.pro.f1": "200 Dokumente / Monat",
  "landing.pricing.pro.f2": "Bis zu 200 Seiten pro Dokument",
-  "landing.pricing.pro.f3": "KI-gestützte Übersetzung",
+  "landing.pricing.pro.f3": "Basis-KI: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google inklusive",
  "landing.pricing.pro.f5": "Individuelle Glossare & Prompts",
  "landing.pricing.pro.f6": "Prioritäts-Support",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "Für Teams mit hohem Bedarf",
  "landing.pricing.business.f1": "1 000 Dokumente / Monat",
  "landing.pricing.business.f2": "Bis zu 500 Seiten pro Dokument",
-  "landing.pricing.business.f3": "Premium-KI (Claude)",
+  "landing.pricing.business.f3": "Premium-KI (Claude 5)",
  "landing.pricing.business.f4": "Alle Anbieter + API-Zugang",
  "landing.pricing.business.f5": "Webhooks & Automatisierung",
  "landing.pricing.business.f6": "5 Teamplätze",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "DOCX, XLSX, PPTX oder PDF Dateien unterstützt",
  "landing.translate.aiAnalysis": "KI-Analyse aktiv",
  "landing.translate.processing": "Verarbeitung läuft",
-  "landing.translate.preservingLayout": "Ihr Layout wird beibehalten"
+  "landing.translate.preservingLayout": "Ihr Layout wird beibehalten",
+  "landing.beforeAfter.seal": "Gleiches Layout, Wort für Wort",
+  "landing.beforeAfter.proof1": "SmartArt-Diagramme werden neu aufgebaut, nicht abgeflacht",
+  "landing.beforeAfter.proof2": "Diagrammreihen und Achsen übersetzt",
+  "landing.beforeAfter.proof3": "Verzeichnisse in der Zielsprache neu erstellt",
+  "landing.beforeAfter.targetTitle": "Technisches Datenblatt — Luftaufbereitung",
+  "landing.beforeAfter.targetPara1": "Die Lüftungsanlage muss vor Ende des zweiten Quartals in Betrieb genommen werden.",
+  "landing.beforeAfter.targetPara2Before": "Die",
+  "landing.beforeAfter.targetTerm": "Lüftungseinheit",
+  "landing.beforeAfter.targetPara2After": "muss gemäß Planung der Filterklasse F7 entsprechen.",
+  "landing.beforeAfter.targetItem1": "Wärmebelastung: 42 kW bei Nenndurchfluss",
+  "landing.beforeAfter.targetItem2": "Schallpegel unter 45 dB(A) bei 3 m",
+  "landing.formats.pill": "KOMPATIBILITÄT",
+  "landing.hero.visualCaption": "Gleiches Layout, neue Sprache — nichts anderes verändert sich",
+  "landing.pricing.free.name": "Kostenlos",
+  "landing.pricing.free.desc": "Ideal zum Entdecken der App",
+  "landing.pricing.free.cta": "Diesen Plan wählen",
+  "landing.pricing.enterprise.name": "Enterprise",
+  "landing.pricing.enterprise.desc": "Maßgeschneiderte Lösungen für große Organisationen",
+  "landing.pricing.enterprise.cta": "Kontakt aufnehmen",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/de/pricing.json
+++ b/frontend/src/lib/i18n/messages/de/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "90 Tage Verlauf",
  "pricing.plans.business.feat1": "1.000 Dokumente / Monat",
  "pricing.plans.business.feat2": "Bis zu 500 Seiten pro Dokument",
-  "pricing.plans.business.feat3": "Basis- + Premium-KI (Claude Haiku)",
+  "pricing.plans.business.feat3": "Basis-KI + Premium-KI (Claude 5)",
  "pricing.plans.business.feat4": "Alle Übersetzungsanbieter",
  "pricing.plans.business.feat5": "Dateien bis zu 50 MB",
  "pricing.plans.business.feat6": "API-Zugang (10.000 Aufrufe/Monat)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "1 Jahr Verlauf",
  "pricing.plans.business.feat10": "Erweiterte Analysen",
  "pricing.plans.enterprise.feat1": "Unbegrenzte Dokumente",
-  "pricing.plans.enterprise.feat2": "Alle KI-Modelle (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "Alle KI-Modelle (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "On-Premise oder dedizierte Cloud",
  "pricing.plans.enterprise.feat4": "99,9 % SLA garantiert",
  "pricing.plans.enterprise.feat5": "24/7 dedizierter Support",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "Ultraschnelle Multi-Thread-KI",
  "pricing.trust.availability.title": "24/7 verfügbar",
  "pricing.trust.availability.sub": "99,9 % garantierte Verfügbarkeit",
-  "pricing.aiModels.title": "Unsere KI-Modelle — März 2026",
+  "pricing.aiModels.title": "Unsere KI-Modelle — September 2026",
  "pricing.aiModels.essential.title": "KI-Basisübersetzung",
  "pricing.aiModels.essential.plan": "Pro-Plan",
  "pricing.aiModels.essential.descPrefix": "Basierend auf",
-  "pricing.aiModels.essential.descSuffix": "— dem kosteneffizientesten KI-Modell 2026. Qualität vergleichbar mit Frontier-Modellen zu einem Bruchteil der Kosten.",
-  "pricing.aiModels.essential.modelName": "unser Essentielles KI-Modell",
-  "pricing.aiModels.essential.context": "163K Token Kontext",
-  "pricing.aiModels.essential.value": "Hervorragendes Preis-Leistungs-Verhältnis",
+  "pricing.aiModels.essential.descSuffix": "— die KI-Modelle mit dem besten Preis-Leistungs-Verhältnis 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash und MiniMax M3",
+  "pricing.aiModels.essential.context": "Bis zu 1,3 M Kontext (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "Das beste Preis-Leistungs-Verhältnis",
  "pricing.aiModels.premium.title": "KI-Premiumübersetzung",
  "pricing.aiModels.premium.plan": "Business-Plan",
  "pricing.aiModels.premium.descPrefix": "Basierend auf",
  "pricing.aiModels.premium.descSuffix": "von Anthropic — präzise bei juristischen, medizinischen und komplexen technischen Dokumenten.",
-  "pricing.aiModels.premium.context": "200K Token Kontext",
+  "pricing.aiModels.premium.context": "1M Kontext",
  "pricing.aiModels.premium.precision": "Höchste Genauigkeit",
  "pricing.faq.title": "Häufig gestellte Fragen",
  "pricing.faq.q1": "Kann ich jederzeit den Plan wechseln?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "Was ist die «KI-Basisübersetzung»?",
  "pricing.faq.a2": "Unser KI-Motor versteht den Kontext Ihrer Dokumente, erhält das Layout und verarbeitet Fachbegriffe deutlich besser als klassische Übersetzungen.",
  "pricing.faq.q3": "Was ist der Unterschied zwischen Basis- und Premium-KI?",
-  "pricing.faq.a3": "Die Basis-KI nutzt ein optimiertes Modell (hervorragendes Preis-Leistungs-Verhältnis). Die Premium-KI verwendet Claude 3.5 Haiku von Anthropic und ist genauer bei juristischen, medizinischen und komplexen technischen Dokumenten.",
+  "pricing.faq.a3": "Die Basis-KI nutzt DeepSeek V4 Flash, GLM-5.3 Flash und MiniMax M3 (hervorragendes Preis-Leistungs-Verhältnis). Die Premium-KI verwendet Claude 5 von Anthropic und ist genauer bei juristischen, medizinischen und komplexen technischen Dokumenten.",
  "pricing.faq.q4": "Werden meine Dokumente nach der Übersetzung gespeichert?",
  "pricing.faq.a4": "Übersetzte Dateien sind je nach Plan verfügbar (30 Tage Starter, 90 Tage Pro, 1 Jahr Business). Sie sind im Ruhezustand und bei der Übertragung verschlüsselt.",
  "pricing.faq.q5": "Was passiert, wenn ich mein monatliches Kontingent überschreite?",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "Fehler beim Erstellen der Zahlung.",
  "pricing.dashboard": "Dashboard",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "Ab 0,09 $ pro Million Token",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Alternativen: DeepSeek V4 Pro und GLM-5.3",
+  "pricing.confirm.title": "Bestätigen Sie Ihr Abonnement",
+  "pricing.confirm.subtitle": "Sie werden zu unserem sicheren Zahlungsanbieter weitergeleitet.",
+  "pricing.confirm.cancel": "Abbrechen",
+  "pricing.confirm.cta": "Weiter zur Zahlung",
+  "pricing.confirm.year": "Jahr",
+  "pricing.confirm.month": "Monat",
+  "pricing.confirm.monthlyEquivalent": "Jährliche Abrechnung — entspricht {price} € / Monat.",
+  "pricing.confirm.secureNote": "Jederzeit über Ihr Profil kündbar. Die Zahlung wird von Stripe abgewickelt; Ihre Kartendaten erreichen nie unsere Server.",
+  "pricing.enterprise.subject": "Enterprise-Angebot",
+  "pricing.enterpriseBand.cta": "Kontakt aufnehmen",
+  "pricing.enterpriseBand.text": "Volumen, dedizierte Engines, On-Premise-Optionen — sprechen wir darüber.",
+  "pricing.error.server": "Serverfehler {status}",
+  "pricing.freeBand.cta": "Kostenlos starten",
+  "pricing.freeBand.text": "Nur ausprobieren? Kostenlos starten — 5 Dokumente pro Monat, ohne Karte.",
+  "pricing.header.titleBase": "Ein Plan für",
+  "pricing.header.titleAccent": "jeden Bedarf",
+  "pricing.toast.close": "Schließen",
+  "pricing.aiModels.premium.price": "2 $ / 10 $ pro 1M Token"
 }
--- a/frontend/src/lib/i18n/messages/de/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/de/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Einheitlicher Zugriff auf die besten Open-Source-Modelle, optimiert für Übersetzung.",
  "providerTheme.openrouter_premium.badge": "Ultra",
  "providerTheme.openrouter_premium.subBadge": "Maximaler Kontext",
-  "providerTheme.openrouter_premium.desc": "Unterstützt durch modernste Modelle (GPT-4o, Claude Sonnet 4.6) für lange Dokumente.",
+  "providerTheme.openrouter_premium.desc": "Unterstützt durch modernste Modelle (GPT-4o, Claude 5) für lange Dokumente.",
  "providerTheme.zai.badge": "Spezialisiert",
  "providerTheme.zai.subBadge": "Finanzen & Recht",
  "providerTheme.zai.desc": "Modell feinabgestimmt auf anspruchsvolle Geschäftsterminologien (Recht, Finanzen).",
--- a/frontend/src/lib/i18n/messages/en/admin.json
+++ b/frontend/src/lib/i18n/messages/en/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Providers",
  "admin.nav.system": "System",
  "admin.nav.logs": "Logs",
+  "admin.nav.stats": "Statistics",
  "admin.users.title": "User Management",
  "admin.users.subtitle": "View and manage user accounts",
  "admin.users.planUpdated": "Plan updated",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "Waiting for data...",
  "admin.system.purging": "Purging...",
  "admin.system.clean": "Clean",
-  "admin.system.purge": "Purge"
+  "admin.system.purge": "Purge",
+  "admin.nav.models": "Models & subscriptions",
+  "admin.nav.marketing": "Marketing",
+  "admin.marketing.audience": "Audience",
+  "admin.marketing.audienceAllUsers": "All accounts",
+  "admin.marketing.audienceInactive": "Accounts inactive for 30 days",
+  "admin.marketing.audiencePlaceholder": "Choose an audience",
+  "admin.marketing.audiencePlanBusiness": "Business plan",
+  "admin.marketing.audiencePlanEnterprise": "Enterprise plan",
+  "admin.marketing.audiencePlanFree": "Free plan",
+  "admin.marketing.audiencePlanPro": "Pro plan",
+  "admin.marketing.audiencePlanStarter": "Starter plan",
+  "admin.marketing.audienceWaitlist": "Waitlist",
+  "admin.marketing.badgeReal": "real",
+  "admin.marketing.badgeTest": "test",
+  "admin.marketing.confirmSend": "Send this email to {count} recipient(s) (“{audience}”)?",
+  "admin.marketing.footerNote": "A footer with the unsubscribe link is automatically added when sending.",
+  "admin.marketing.hidePreview": "Hide preview",
+  "admin.marketing.history": "Send history",
+  "admin.marketing.historyFailedSuffix": ", {count} failed",
+  "admin.marketing.historySent": "{sent}/{total} sent",
+  "admin.marketing.html": "HTML content",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Please fill in the subject and the HTML.",
+  "admin.marketing.incompleteTitle": "Incomplete content",
+  "admin.marketing.loadErrorTitle": "Loading error",
+  "admin.marketing.localCheckUnavailable": "Local check unavailable in this browser: the server will verify the prior test send.",
+  "admin.marketing.networkDesc": "Cannot reach the backend.",
+  "admin.marketing.newCampaign": "New campaign",
+  "admin.marketing.noHistory": "No send recorded.",
+  "admin.marketing.preview": "Preview",
+  "admin.marketing.previewEmpty": "Preview: paste your HTML next to this panel.",
+  "admin.marketing.previewTitle": "Email preview",
+  "admin.marketing.recipientsAvailable": "{count} recipient(s) available after excluding unsubscribed contacts.",
+  "admin.marketing.sendReal": "Send to {count} recipient(s)",
+  "admin.marketing.sendRefusedTitle": "Send refused",
+  "admin.marketing.sendTest": "Send test",
+  "admin.marketing.sendingDesc": "{count} email(s) queued for sending (0.2 s interval). Unsubscribed contacts are excluded.",
+  "admin.marketing.sendingTitle": "Send in progress",
+  "admin.marketing.subject": "Subject",
+  "admin.marketing.subjectPlaceholder": "E.g.: Your translation is waiting — 20% off for 7 days",
+  "admin.marketing.subtitle": "Send emails to an audience, with a mandatory test send, automatic unsubscribe link and full history.",
+  "admin.marketing.testDoneDesc": "Email received at {email}",
+  "admin.marketing.testDoneTitle": "Test send done",
+  "admin.marketing.testEmail": "Test email (optional)",
+  "admin.marketing.testEmailPlaceholder": "Otherwise: SMTP sender address",
+  "admin.marketing.testFailedTitle": "Test send failed",
+  "admin.marketing.testRequiredHint": "A real send requires a prior test send of this exact content (verified by the server).",
+  "admin.marketing.testValidated": "Test send validated for this content.",
+  "admin.marketing.title": "Marketing — email campaigns",
+  "admin.marketing.unsubscribedNote": "{count} unsubscribed contact(s) will be excluded from any send.",
+  "admin.models.addModel": "Add a model",
+  "admin.models.businessBadge": "Business plan",
+  "admin.models.cancel": "Cancel",
+  "admin.models.catalog": "OpenRouter catalog",
+  "admin.models.catalogErrorDesc": "Failed to load the OpenRouter catalog.",
+  "admin.models.catalogErrorTitle": "Catalog unavailable",
+  "admin.models.customBadge": "Custom tier",
+  "admin.models.customNote": "Bespoke models, agreed case by case with the customer.",
+  "admin.models.defaultBadge": "default",
+  "admin.models.emptyTier": "No model: the plan's official range will be used.",
+  "admin.models.enterpriseBadge": "Enterprise plan",
+  "admin.models.essentialBadge": "Essential AI tier",
+  "admin.models.essentialCost": "— billed at cost factor 1",
+  "admin.models.fallbackFirst": "(fallback #1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — check your admin token.",
+  "admin.models.loadErrorTitle": "Loading error",
+  "admin.models.matrixDesc": "Active models per tier — the list order is the fallback priority. The radio button picks the default model, applied immediately without redeployment.",
+  "admin.models.matrixTitle": "Model matrix by plan",
+  "admin.models.missingConfig": "AI tier configuration not found. Reload the page.",
+  "admin.models.moveDown": "Move down",
+  "admin.models.moveUp": "Move up",
+  "admin.models.networkDesc": "Cannot reach the backend.",
+  "admin.models.networkTitle": "Network error",
+  "admin.models.premiumBadge": "Premium AI tier",
+  "admin.models.premiumCost": "— billed at cost factor 5",
+  "admin.models.premiumReservedNote": "Reserved for Business and Enterprise plans (“openrouter_premium” engine).",
+  "admin.models.proBadge": "Pro plan",
+  "admin.models.remove": "Remove",
+  "admin.models.save": "Save",
+  "admin.models.saveErrorTitle": "Save error",
+  "admin.models.saveNetworkDesc": "Could not save the configuration.",
+  "admin.models.savedDesc": "The new default model applies from the next translation, without redeployment.",
+  "admin.models.savedTitle": "Models saved",
+  "admin.models.saving": "Saving...",
+  "admin.models.setDefault": "Set {model} as default model",
+  "admin.models.sharedTierNote": "Shared tier: the Business plan also uses the Essential tier for its “openrouter” engine — the list above applies to both plans.",
+  "admin.models.subtitle": "Plans → AI tiers matrix. The model actually used per translation is the plan's tier model: a Pro plan can never trigger a Premium model.",
+  "admin.models.title": "Models & subscriptions",
+  "admin.stats.aiTiers": "AI tiers (30 d)",
+  "admin.stats.includingCredits": "including {amount} in credits",
+  "admin.stats.mrr": "Estimated MRR",
+  "admin.stats.noTranslationsYet": "No translation stored in the last 30 days: the AI tier breakdown will fill in with upcoming translations.",
+  "admin.stats.payments30": "{count} payment(s) in 30 days",
+  "admin.stats.refreshing": "Refreshing...",
+  "admin.stats.revenue30": "Revenue (30 days)",
+  "admin.stats.revenueTotal": "Total collected revenue",
+  "admin.stats.tiersSub": "Essential / Premium · classic {classic} · other {other}",
+  "admin.stats.unavailable": "Business statistics unavailable ({error}).",
+  "admin.stats.waitlist": "Waitlist",
+  "admin.stats.waitlistSub": "people waiting"
 }
--- a/frontend/src/lib/i18n/messages/en/landing.json
+++ b/frontend/src/lib/i18n/messages/en/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "For demanding professionals",
  "landing.pricing.pro.f1": "200 documents / month",
  "landing.pricing.pro.f2": "Up to 200 pages per doc",
-  "landing.pricing.pro.f3": "AI-powered translation",
+  "landing.pricing.pro.f3": "Essential AI: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google included",
  "landing.pricing.pro.f5": "Custom glossaries & prompts",
  "landing.pricing.pro.f6": "Priority support",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "For teams with high-volume needs",
  "landing.pricing.business.f1": "1,000 documents / month",
  "landing.pricing.business.f2": "Up to 500 pages per doc",
-  "landing.pricing.business.f3": "Premium AI (Claude)",
+  "landing.pricing.business.f3": "Premium AI (Claude 5)",
  "landing.pricing.business.f4": "All providers + API access",
  "landing.pricing.business.f5": "Webhooks & automation",
  "landing.pricing.business.f6": "5 team seats",
--- a/frontend/src/lib/i18n/messages/en/pricing.json
+++ b/frontend/src/lib/i18n/messages/en/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "90-day history",
  "pricing.plans.business.feat1": "1,000 documents / month",
  "pricing.plans.business.feat2": "Up to 500 pages per document",
-  "pricing.plans.business.feat3": "Essential + Premium AI (Claude Haiku)",
+  "pricing.plans.business.feat3": "Essential AI + Premium AI (Claude 5)",
  "pricing.plans.business.feat4": "All translation providers",
  "pricing.plans.business.feat5": "Files up to 50 MB",
  "pricing.plans.business.feat6": "API access (10,000 calls/month)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "1-year history",
  "pricing.plans.business.feat10": "Advanced analytics",
  "pricing.plans.enterprise.feat1": "Unlimited documents",
-  "pricing.plans.enterprise.feat2": "All AI models (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "All AI models (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "On-premise or dedicated cloud deployment",
  "pricing.plans.enterprise.feat4": "99.9% SLA guaranteed",
  "pricing.plans.enterprise.feat5": "24/7 dedicated support",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "Ultra-fast multi-threaded AI",
  "pricing.trust.availability.title": "Available 24/7",
  "pricing.trust.availability.sub": "99.9% guaranteed uptime",
-  "pricing.aiModels.title": "Our AI Models — March 2026",
+  "pricing.aiModels.title": "Our AI Models — September 2026",
  "pricing.aiModels.essential.title": "Essential AI Translation",
  "pricing.aiModels.essential.plan": "Pro Plan",
  "pricing.aiModels.essential.descPrefix": "Based on",
-  "pricing.aiModels.essential.descSuffix": "— the most cost-effective AI model of 2026. Quality comparable to frontier models at a fraction of the cost.",
-  "pricing.aiModels.essential.modelName": "our Essential AI model",
-  "pricing.aiModels.essential.context": "163K tokens of context",
-  "pricing.aiModels.essential.value": "Excellent value for money",
+  "pricing.aiModels.essential.descSuffix": "— the most cost-effective AI models of 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash and MiniMax M3",
+  "pricing.aiModels.essential.context": "Up to 1.3M context (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "The best value for money",
  "pricing.aiModels.premium.title": "Premium AI Translation",
  "pricing.aiModels.premium.plan": "Business Plan",
  "pricing.aiModels.premium.descPrefix": "Based on",
  "pricing.aiModels.premium.descSuffix": "by Anthropic — accurate on legal, medical and complex technical documents.",
-  "pricing.aiModels.premium.context": "200K tokens of context",
+  "pricing.aiModels.premium.context": "1M context",
  "pricing.aiModels.premium.precision": "Best accuracy",
  "pricing.faq.title": "Frequently asked questions",
  "pricing.faq.q1": "Can I change plans at any time?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "What is \\\"Essential AI Translation\\\"?",
  "pricing.faq.a2": "It's our AI engine. It understands your documents' context, preserves layout and handles technical terms much better than classic translation.",
  "pricing.faq.q3": "What's the difference between Essential and Premium AI?",
-  "pricing.faq.a3": "Essential AI uses an optimized model (excellent value for money). Premium AI uses Anthropic's Claude Sonnet 4.6, more accurate on legal, medical and complex technical documents.",
+  "pricing.faq.a3": "Essential AI runs on DeepSeek V4 Flash, GLM-5.3 Flash and MiniMax M3 (excellent value for money). Premium AI uses Anthropic's Claude 5, more accurate on legal, medical and complex technical documents.",
  "pricing.faq.q4": "Are my documents kept after translation?",
  "pricing.faq.a4": "Translated files are available according to your plan (30 days Starter, 90 days Pro, 1 year Business). They are encrypted at rest and in transit.",
  "pricing.faq.q5": "What happens if I exceed my monthly quota?",
@@ -163,5 +163,9 @@
  "pricing.freeBand.text": "Just want to try? Start free — 5 documents per month, no card required.",
  "pricing.freeBand.cta": "Start free",
  "pricing.enterpriseBand.text": "Volume, dedicated engines, on-premise options — let's talk.",
-  "pricing.enterpriseBand.cta": "Contact us"
+  "pricing.enterpriseBand.cta": "Contact us",
+  "pricing.aiModels.essential.price": "From $0.09 per million tokens",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Alternatives: DeepSeek V4 Pro and GLM-5.3",
+  "pricing.aiModels.premium.price": "$2 / $10 per 1M tokens"
 }
--- a/frontend/src/lib/i18n/messages/en/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/en/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Unified access to the best open-source models optimized for translation.",
  "providerTheme.openrouter_premium.badge": "Ultra",
  "providerTheme.openrouter_premium.subBadge": "Maximum context",
-  "providerTheme.openrouter_premium.desc": "Assisted by state-of-the-art models (GPT-4o, Claude Sonnet 4.6) for long documents.",
+  "providerTheme.openrouter_premium.desc": "Assisted by state-of-the-art models (GPT-4o, Claude 5) for long documents.",
  "providerTheme.zai.badge": "Specialized",
  "providerTheme.zai.subBadge": "Finance & Law",
  "providerTheme.zai.desc": "Model fine-tuned for demanding business terminologies (legal, finance).",
--- a/frontend/src/lib/i18n/messages/es/admin.json
+++ b/frontend/src/lib/i18n/messages/es/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Proveedores",
  "admin.nav.system": "Sistema",
  "admin.nav.logs": "Registros",
+  "admin.nav.stats": "Estadísticas",
  "admin.users.title": "Gestión de Usuarios",
  "admin.users.subtitle": "Ver y gestionar cuentas de usuario",
  "admin.users.planUpdated": "Plan actualizado",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "Esperando datos...",
  "admin.system.purging": "Purgando...",
  "admin.system.clean": "Limpiar",
-  "admin.system.purge": "Purgar"
+  "admin.system.purge": "Purgar",
+  "admin.nav.models": "Modelos y suscripciones",
+  "admin.nav.marketing": "Marketing",
+  "admin.marketing.audience": "Audiencia",
+  "admin.marketing.audienceAllUsers": "Todas las cuentas",
+  "admin.marketing.audienceInactive": "Cuentas inactivas desde hace 30 días",
+  "admin.marketing.audiencePlaceholder": "Elegir una audiencia",
+  "admin.marketing.audiencePlanBusiness": "Plan Business",
+  "admin.marketing.audiencePlanEnterprise": "Plan Enterprise",
+  "admin.marketing.audiencePlanFree": "Plan Gratuito",
+  "admin.marketing.audiencePlanPro": "Plan Pro",
+  "admin.marketing.audiencePlanStarter": "Plan Starter",
+  "admin.marketing.audienceWaitlist": "Lista de espera",
+  "admin.marketing.badgeReal": "real",
+  "admin.marketing.badgeTest": "prueba",
+  "admin.marketing.confirmSend": "¿Enviar este correo a {count} destinatario(s) («{audience}»)?",
+  "admin.marketing.footerNote": "Un pie de página con el enlace de baja se añade automáticamente al enviar.",
+  "admin.marketing.hidePreview": "Ocultar vista previa",
+  "admin.marketing.history": "Historial de envíos",
+  "admin.marketing.historyFailedSuffix": ", {count} fallos",
+  "admin.marketing.historySent": "{sent}/{total} enviados",
+  "admin.marketing.html": "Contenido HTML",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Rellena el asunto y el HTML.",
+  "admin.marketing.incompleteTitle": "Contenido incompleto",
+  "admin.marketing.loadErrorTitle": "Error de carga",
+  "admin.marketing.localCheckUnavailable": "Verificación local no disponible en este navegador: el servidor controlará el envío de prueba previo.",
+  "admin.marketing.networkDesc": "No se puede contactar con el backend.",
+  "admin.marketing.newCampaign": "Nuevo envío",
+  "admin.marketing.noHistory": "Ningún envío registrado.",
+  "admin.marketing.preview": "Vista previa",
+  "admin.marketing.previewEmpty": "Vista previa: pega tu HTML al lado.",
+  "admin.marketing.previewTitle": "Vista previa del correo",
+  "admin.marketing.recipientsAvailable": "{count} destinatario(s) disponibles tras excluir las bajas.",
+  "admin.marketing.sendReal": "Enviar a {count} destinatario(s)",
+  "admin.marketing.sendRefusedTitle": "Envío rechazado",
+  "admin.marketing.sendTest": "Envío de prueba",
+  "admin.marketing.sendingDesc": "{count} correo(s) en cola de envío (intervalo de 0,2 s). Las bajas quedan excluidas.",
+  "admin.marketing.sendingTitle": "Envío en curso",
+  "admin.marketing.subject": "Asunto",
+  "admin.marketing.subjectPlaceholder": "Ej.: Tu traducción te espera — 20 % durante 7 días",
+  "admin.marketing.subtitle": "Envío de correos a una audiencia, con envío de prueba obligatorio, enlace de baja automático e historial completo.",
+  "admin.marketing.testDoneDesc": "Correo recibido en {email}",
+  "admin.marketing.testDoneTitle": "Envío de prueba realizado",
+  "admin.marketing.testEmail": "Correo de prueba (opcional)",
+  "admin.marketing.testEmailPlaceholder": "Si no: dirección remitente SMTP",
+  "admin.marketing.testFailedTitle": "Fallo del envío de prueba",
+  "admin.marketing.testRequiredHint": "El envío real exige un envío de prueba previo de este contenido exacto (verificado por el servidor).",
+  "admin.marketing.testValidated": "Envío de prueba validado para este contenido.",
+  "admin.marketing.title": "Marketing — envíos de correo",
+  "admin.marketing.unsubscribedNote": "{count} contacto(s) de baja serán excluidos de todo envío.",
+  "admin.models.addModel": "Añadir un modelo",
+  "admin.models.businessBadge": "Plan Business",
+  "admin.models.cancel": "Cancelar",
+  "admin.models.catalog": "Catálogo de OpenRouter",
+  "admin.models.catalogErrorDesc": "No se pudo cargar el catálogo de OpenRouter.",
+  "admin.models.catalogErrorTitle": "Catálogo no disponible",
+  "admin.models.customBadge": "Nivel personalizado",
+  "admin.models.customNote": "Modelos a medida, definidos caso por caso con el cliente.",
+  "admin.models.defaultBadge": "por defecto",
+  "admin.models.emptyTier": "Sin modelos: se usará la gama oficial del plan.",
+  "admin.models.enterpriseBadge": "Plan Enterprise",
+  "admin.models.essentialBadge": "Nivel de IA Esencial",
+  "admin.models.essentialCost": "— facturado con factor de coste 1",
+  "admin.models.fallbackFirst": "(respaldo n.º 1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — comprueba tu token de administrador.",
+  "admin.models.loadErrorTitle": "Error de carga",
+  "admin.models.matrixDesc": "Modelos activos por nivel — el orden de la lista es la prioridad de respaldo. El botón de opción elige el modelo por defecto, aplicado al instante sin redespliegue.",
+  "admin.models.matrixTitle": "Matriz de modelos por plan",
+  "admin.models.missingConfig": "Configuración de niveles de IA no encontrada. Recarga la página.",
+  "admin.models.moveDown": "Bajar",
+  "admin.models.moveUp": "Subir",
+  "admin.models.networkDesc": "No se puede contactar con el backend.",
+  "admin.models.networkTitle": "Error de red",
+  "admin.models.premiumBadge": "Nivel de IA Premium",
+  "admin.models.premiumCost": "— facturado con factor de coste 5",
+  "admin.models.premiumReservedNote": "Reservado a los planes Business y Enterprise (motor «openrouter_premium»).",
+  "admin.models.proBadge": "Plan Pro",
+  "admin.models.remove": "Quitar",
+  "admin.models.save": "Guardar",
+  "admin.models.saveErrorTitle": "Error al guardar",
+  "admin.models.saveNetworkDesc": "No se pudo guardar la configuración.",
+  "admin.models.savedDesc": "El nuevo modelo por defecto se usa desde la siguiente traducción, sin redespliegue.",
+  "admin.models.savedTitle": "Modelos guardados",
+  "admin.models.saving": "Guardando...",
+  "admin.models.setDefault": "Definir {model} como modelo por defecto",
+  "admin.models.sharedTierNote": "Nivel compartido: el plan Business también usa el nivel Esencial con su motor «openrouter» — la lista de arriba se aplica a ambos planes.",
+  "admin.models.subtitle": "Matriz planes → niveles de IA. El modelo realmente usado en cada traducción es el del nivel del plan: un plan Pro nunca puede disparar un modelo Premium.",
+  "admin.models.title": "Modelos y suscripciones",
+  "admin.stats.aiTiers": "Niveles de IA (30 d)",
+  "admin.stats.includingCredits": "incluidos {amount} en créditos",
+  "admin.stats.mrr": "MRR estimado",
+  "admin.stats.noTranslationsYet": "Sin traducciones registradas en los últimos 30 días: el reparto por niveles se completará con las próximas traducciones.",
+  "admin.stats.payments30": "{count} pago(s) en 30 días",
+  "admin.stats.refreshing": "Actualizando...",
+  "admin.stats.revenue30": "Ingresos (30 días)",
+  "admin.stats.revenueTotal": "Ingresos cobrados (total)",
+  "admin.stats.tiersSub": "Esencial / Premium · clásico {classic} · otro {other}",
+  "admin.stats.unavailable": "Estadísticas de negocio no disponibles ({error}).",
+  "admin.stats.waitlist": "Lista de espera",
+  "admin.stats.waitlistSub": "inscritos en espera"
 }
--- a/frontend/src/lib/i18n/messages/es/landing.json
+++ b/frontend/src/lib/i18n/messages/es/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "Para profesionales exigentes",
  "landing.pricing.pro.f1": "200 documentos / mes",
  "landing.pricing.pro.f2": "Hasta 200 páginas por doc",
-  "landing.pricing.pro.f3": "Traducción con IA",
+  "landing.pricing.pro.f3": "IA Esencial: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google incluidos",
  "landing.pricing.pro.f5": "Glosarios y prompts personalizados",
  "landing.pricing.pro.f6": "Soporte prioritario",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "Para equipos con grandes necesidades",
  "landing.pricing.business.f1": "1 000 documentos / mes",
  "landing.pricing.business.f2": "Hasta 500 páginas por doc",
-  "landing.pricing.business.f3": "IA Premium (Claude)",
+  "landing.pricing.business.f3": "IA Premium (Claude 5)",
  "landing.pricing.business.f4": "Todos los proveedores + API",
  "landing.pricing.business.f5": "Webhooks y automatización",
  "landing.pricing.business.f6": "5 puestos de equipo",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "Archivos DOCX, XLSX, PPTX o PDF compatibles",
  "landing.translate.aiAnalysis": "Análisis IA Activo",
  "landing.translate.processing": "Procesando",
-  "landing.translate.preservingLayout": "Tu diseño se está preservando"
+  "landing.translate.preservingLayout": "Tu diseño se está preservando",
+  "landing.beforeAfter.seal": "El mismo diseño, palabra por palabra",
+  "landing.beforeAfter.proof1": "Los diagramas SmartArt se reconstruyen, no se aplanan",
+  "landing.beforeAfter.proof2": "Series y ejes de los gráficos traducidos",
+  "landing.beforeAfter.proof3": "Índices regenerados en el idioma de destino",
+  "landing.beforeAfter.targetTitle": "Pliego técnico — Tratamiento de aire",
+  "landing.beforeAfter.targetPara1": "La instalación de ventilación debe ponerse en servicio antes de finales del segundo trimestre.",
+  "landing.beforeAfter.targetPara2Before": "La",
+  "landing.beforeAfter.targetTerm": "unidad de tratamiento de aire",
+  "landing.beforeAfter.targetPara2After": "debe cumplir la clase de filtro F7 según el calendario.",
+  "landing.beforeAfter.targetItem1": "Carga térmica: 42 kW a caudal nominal",
+  "landing.beforeAfter.targetItem2": "Nivel sonoro inferior a 45 dB(A) a 3 m",
+  "landing.formats.pill": "COMPATIBILIDAD",
+  "landing.hero.visualCaption": "El mismo diseño, otro idioma — nada más cambia",
+  "landing.pricing.free.name": "Gratis",
+  "landing.pricing.free.desc": "Ideal para descubrir la aplicación",
+  "landing.pricing.free.cta": "Elegir este plan",
+  "landing.pricing.enterprise.name": "Empresas",
+  "landing.pricing.enterprise.desc": "Soluciones a medida para grandes organizaciones",
+  "landing.pricing.enterprise.cta": "Contáctenos",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/es/pricing.json
+++ b/frontend/src/lib/i18n/messages/es/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "Historial de 90 días",
  "pricing.plans.business.feat1": "1 000 documentos / mes",
  "pricing.plans.business.feat2": "Hasta 500 páginas por documento",
-  "pricing.plans.business.feat3": "IA Esencial + Premium (Claude Haiku)",
+  "pricing.plans.business.feat3": "IA Esencial + IA Premium (Claude 5)",
  "pricing.plans.business.feat4": "Todos los proveedores de traducción",
  "pricing.plans.business.feat5": "Archivos de hasta 50 MB",
  "pricing.plans.business.feat6": "Acceso API (10 000 llamadas/mes)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "Historial de 1 año",
  "pricing.plans.business.feat10": "Analíticas avanzadas",
  "pricing.plans.enterprise.feat1": "Documentos ilimitados",
-  "pricing.plans.enterprise.feat2": "Todos los modelos de IA (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "Todos los modelos de IA (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "Implementación local o en nube dedicada",
  "pricing.plans.enterprise.feat4": "SLA 99,9 % garantizado",
  "pricing.plans.enterprise.feat5": "Soporte dedicado 24/7",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "IA multiproceso ultrarrápida",
  "pricing.trust.availability.title": "Disponible 24/7",
  "pricing.trust.availability.sub": "Disponibilidad garantizada del 99,9 %",
-  "pricing.aiModels.title": "Nuestros modelos de IA — Marzo 2026",
+  "pricing.aiModels.title": "Nuestros modelos de IA — Septiembre 2026",
  "pricing.aiModels.essential.title": "Traducción IA Esencial",
  "pricing.aiModels.essential.plan": "Plan Pro",
  "pricing.aiModels.essential.descPrefix": "Basado en",
-  "pricing.aiModels.essential.descSuffix": "— el modelo de IA más rentable de 2026. Calidad comparable a los modelos frontier a una fracción del coste.",
-  "pricing.aiModels.essential.modelName": "nuestro modelo IA Esencial",
-  "pricing.aiModels.essential.context": "163K tokens de contexto",
-  "pricing.aiModels.essential.value": "Excelente relación calidad-precio",
+  "pricing.aiModels.essential.descSuffix": "— los modelos de IA con la mejor relación calidad-precio de 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash y MiniMax M3",
+  "pricing.aiModels.essential.context": "Hasta 1,3 M de contexto (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "La mejor relación calidad-precio",
  "pricing.aiModels.premium.title": "Traducción IA Premium",
  "pricing.aiModels.premium.plan": "Plan Business",
  "pricing.aiModels.premium.descPrefix": "Basado en",
  "pricing.aiModels.premium.descSuffix": "de Anthropic — preciso en documentos jurídicos, médicos y técnicos complejos.",
-  "pricing.aiModels.premium.context": "200K tokens de contexto",
+  "pricing.aiModels.premium.context": "1M de contexto",
  "pricing.aiModels.premium.precision": "Máxima precisión",
  "pricing.faq.title": "Preguntas frecuentes",
  "pricing.faq.q1": "¿Puedo cambiar de plan en cualquier momento?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "¿Qué es la «Traducción IA Esencial»?",
  "pricing.faq.a2": "Es nuestro motor de IA. Comprende el contexto de tus documentos, conserva el diseño y maneja términos técnicos mucho mejor que una traducción clásica.",
  "pricing.faq.q3": "¿Cuál es la diferencia entre IA Esencial e IA Premium?",
-  "pricing.faq.a3": "La IA Esencial usa un modelo optimizado (excelente relación calidad-precio). La IA Premium usa Claude 3.5 Haiku de Anthropic, más precisa en documentos jurídicos, médicos y técnicos complejos.",
+  "pricing.faq.a3": "La IA Esencial se apoya en DeepSeek V4 Flash, GLM-5.3 Flash y MiniMax M3 (excelente relación calidad-precio). La IA Premium usa Claude 5 de Anthropic, más precisa en documentos jurídicos, médicos y técnicos complejos.",
  "pricing.faq.q4": "¿Se conservan mis documentos después de la traducción?",
  "pricing.faq.a4": "Los archivos traducidos están disponibles según tu plan (30 días Starter, 90 días Pro, 1 año Business). Están cifrados en reposo y en tránsito.",
  "pricing.faq.q5": "¿Qué ocurre si supero mi cuota mensual?",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "Error al crear el pago.",
  "pricing.dashboard": "Panel",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "Desde 0,09 $ por millón de tokens",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Alternativas: DeepSeek V4 Pro y GLM-5.3",
+  "pricing.confirm.title": "Confirma tu suscripción",
+  "pricing.confirm.subtitle": "Serás redirigido a nuestro proveedor de pago seguro.",
+  "pricing.confirm.cancel": "Cancelar",
+  "pricing.confirm.cta": "Continuar al pago",
+  "pricing.confirm.year": "año",
+  "pricing.confirm.month": "mes",
+  "pricing.confirm.monthlyEquivalent": "Facturación anual — equivalente a {price} € / mes.",
+  "pricing.confirm.secureNote": "Cancelable en cualquier momento desde tu perfil. El pago lo procesa Stripe; los datos de tu tarjeta nunca pasan por nuestros servidores.",
+  "pricing.enterprise.subject": "Consulta Enterprise",
+  "pricing.enterpriseBand.cta": "Contáctenos",
+  "pricing.enterpriseBand.text": "Volumen, motores dedicados, opciones on-premise — hablemos.",
+  "pricing.error.server": "Error del servidor {status}",
+  "pricing.freeBand.cta": "Empieza gratis",
+  "pricing.freeBand.text": "¿Solo quieres probar? Empieza gratis — 5 documentos al mes, sin tarjeta.",
+  "pricing.header.titleBase": "Un plan para",
+  "pricing.header.titleAccent": "cada necesidad",
+  "pricing.toast.close": "Cerrar",
+  "pricing.aiModels.premium.price": "2 $ / 10 $ por 1M de tokens"
 }
--- a/frontend/src/lib/i18n/messages/es/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/es/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Acceso unificado a los mejores modelos de código abierto optimizados para traducción.",
  "providerTheme.openrouter_premium.badge": "Ultra",
  "providerTheme.openrouter_premium.subBadge": "Contexto máximo",
-  "providerTheme.openrouter_premium.desc": "Asistido por modelos de última generación (GPT-4o, Claude Sonnet 4.6) para documentos largos.",
+  "providerTheme.openrouter_premium.desc": "Asistido por modelos de última generación (GPT-4o, Claude 5) para documentos largos.",
  "providerTheme.zai.badge": "Especializada",
  "providerTheme.zai.subBadge": "Finanzas y Derecho",
  "providerTheme.zai.desc": "Modelo ajustado para terminologías empresariales exigentes (legal, finanzas).",
--- a/frontend/src/lib/i18n/messages/fa/admin.json
+++ b/frontend/src/lib/i18n/messages/fa/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "ارائه‌دهندگان",
  "admin.nav.system": "سیستم",
  "admin.nav.logs": "گزارش‌ها",
+  "admin.nav.stats": "آمار",
  "admin.users.title": "مدیریت کاربران",
  "admin.users.subtitle": "مشاهده و مدیریت حساب‌های کاربری",
  "admin.users.planUpdated": "طرح به‌روز شد",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "در انتظار داده...",
  "admin.system.purging": "در حال پاکسازی...",
  "admin.system.clean": "پاکسازی",
-  "admin.system.purge": "حذف کامل"
+  "admin.system.purge": "حذف کامل",
+  "admin.nav.models": "مدل‌ها و اشتراک‌ها",
+  "admin.nav.marketing": "بازاریابی",
+  "admin.marketing.audience": "مخاطب",
+  "admin.marketing.audienceAllUsers": "همه حساب‌ها",
+  "admin.marketing.audienceInactive": "حساب‌های غیرفعال از ۳۰ روز پیش",
+  "admin.marketing.audiencePlaceholder": "انتخاب مخاطب",
+  "admin.marketing.audiencePlanBusiness": "اشتراک Business",
+  "admin.marketing.audiencePlanEnterprise": "اشتراک Enterprise",
+  "admin.marketing.audiencePlanFree": "اشتراک رایگان",
+  "admin.marketing.audiencePlanPro": "اشتراک Pro",
+  "admin.marketing.audiencePlanStarter": "اشتراک Starter",
+  "admin.marketing.audienceWaitlist": "فهرست انتظار",
+  "admin.marketing.badgeReal": "واقعی",
+  "admin.marketing.badgeTest": "آزمایشی",
+  "admin.marketing.confirmSend": "این ایمیل به {count} گیرنده ارسال شود («{audience}»)؟",
+  "admin.marketing.footerNote": "یک پانوشت با پیوند لغو اشتراک به‌طور خودکار به ارسال افزوده می‌شود.",
+  "admin.marketing.hidePreview": "پنهان کردن پیش‌نمایش",
+  "admin.marketing.history": "تاریخچه ارسال‌ها",
+  "admin.marketing.historyFailedSuffix": "، {count} ناموفق",
+  "admin.marketing.historySent": "{sent}/{total} ارسال‌شده",
+  "admin.marketing.html": "محتوای HTML",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "موضوع و HTML را پر کنید.",
+  "admin.marketing.incompleteTitle": "محتوای ناقص",
+  "admin.marketing.loadErrorTitle": "خطا در بارگذاری",
+  "admin.marketing.localCheckUnavailable": "بررسی محلی در این مرورگر در دسترس نیست: سرور ارسال آزمایشی قبلی را کنترل خواهد کرد.",
+  "admin.marketing.networkDesc": "دسترسی به بک‌اند ممکن نیست.",
+  "admin.marketing.newCampaign": "کمپین جدید",
+  "admin.marketing.noHistory": "هیچ ارسالی ثبت نشده است.",
+  "admin.marketing.preview": "پیش‌نمایش",
+  "admin.marketing.previewEmpty": "پیش‌نمایش: HTML خود را در کنار این بوم بچسبانید.",
+  "admin.marketing.previewTitle": "پیش‌نمایش ایمیل",
+  "admin.marketing.recipientsAvailable": "{count} گیرنده پس از کنارگذاشتن لغواشتراک‌ها در دسترس است.",
+  "admin.marketing.sendReal": "ارسال به {count} گیرنده",
+  "admin.marketing.sendRefusedTitle": "ارسال رد شد",
+  "admin.marketing.sendTest": "ارسال آزمایشی",
+  "admin.marketing.sendingDesc": "{count} ایمیل در صف ارسال (فاصله ۰٫۲ ثانیه). لغواشتراک‌ها مستثنی هستند.",
+  "admin.marketing.sendingTitle": "ارسال در جریان است",
+  "admin.marketing.subject": "موضوع",
+  "admin.marketing.subjectPlaceholder": "مثال: ترجمه شما منتظر است — ۲۰٪ به مدت ۷ روز",
+  "admin.marketing.subtitle": "ارسال ایمیل به یک مخاطب، با ارسال آزمایشی الزامی، پیوند لغو اشتراک خودکار و تاریخچه کامل.",
+  "admin.marketing.testDoneDesc": "ایمیل به {email} رسید",
+  "admin.marketing.testDoneTitle": "ارسال آزمایشی انجام شد",
+  "admin.marketing.testEmail": "ایمیل آزمایشی (اختیاری)",
+  "admin.marketing.testEmailPlaceholder": "در غیر این صورت: آدرس فرستنده SMTP",
+  "admin.marketing.testFailedTitle": "ارسال آزمایشی ناموفق بود",
+  "admin.marketing.testRequiredHint": "ارسال واقعی نیازمند ارسال آزمایشی قبلی همین محتوا است (تأییدشده توسط سرور).",
+  "admin.marketing.testValidated": "ارسال آزمایشی این محتوا تأیید شده است.",
+  "admin.marketing.title": "بازاریابی — کمپین‌های ایمیلی",
+  "admin.marketing.unsubscribedNote": "{count} لغواشتراک از هر ارسالی مستثنی خواهند شد.",
+  "admin.models.addModel": "افزودن مدل",
+  "admin.models.businessBadge": "اشتراک Business",
+  "admin.models.cancel": "لغو",
+  "admin.models.catalog": "کاتالوگ OpenRouter",
+  "admin.models.catalogErrorDesc": "بارگذاری کاتالوگ OpenRouter ممکن نشد.",
+  "admin.models.catalogErrorTitle": "کاتالوگ در دسترس نیست",
+  "admin.models.customBadge": "سطح سفارشی",
+  "admin.models.customNote": "مدل‌های سفارشی، مورد به مورد با مشتری توافق می‌شود.",
+  "admin.models.defaultBadge": "پیش‌فرض",
+  "admin.models.emptyTier": "بدون مدل: مجموعه رسمی اشتراک استفاده خواهد شد.",
+  "admin.models.enterpriseBadge": "اشتراک Enterprise",
+  "admin.models.essentialBadge": "سطح هوش مصنوعی پایه",
+  "admin.models.essentialCost": "— محاسبه با ضریب هزینه ۱",
+  "admin.models.fallbackFirst": "(پشتیبان شماره ۱)",
+  "admin.models.loadErrorDesc": "HTTP {status} — توکن مدیر را بررسی کنید.",
+  "admin.models.loadErrorTitle": "خطا در بارگذاری",
+  "admin.models.matrixDesc": "مدل‌های فعال هر سطح — ترتیب فهرست، اولویت پشتیبان است. دکمه رادیویی مدل پیش‌فرض را انتخاب می‌کند که بلافاصله و بدون استقرار مجدد اعمال می‌شود.",
+  "admin.models.matrixTitle": "ماتریس مدل‌ها به تفکیک اشتراک",
+  "admin.models.missingConfig": "پیکربندی سطوح هوش مصنوعی یافت نشد. صفحه را دوباره بارگذاری کنید.",
+  "admin.models.moveDown": "انتقال به پایین",
+  "admin.models.moveUp": "انتقال به بالا",
+  "admin.models.networkDesc": "دسترسی به بک‌اند ممکن نیست.",
+  "admin.models.networkTitle": "خطای شبکه",
+  "admin.models.premiumBadge": "سطح هوش مصنوعی پیشرفته",
+  "admin.models.premiumCost": "— محاسبه با ضریب هزینه ۵",
+  "admin.models.premiumReservedNote": "مخصوص اشتراک‌های Business و Enterprise (موتور «openrouter_premium»).",
+  "admin.models.proBadge": "اشتراک Pro",
+  "admin.models.remove": "حذف",
+  "admin.models.save": "ذخیره",
+  "admin.models.saveErrorTitle": "خطا در ذخیره",
+  "admin.models.saveNetworkDesc": "ذخیره پیکربندی ممکن نشد.",
+  "admin.models.savedDesc": "مدل پیش‌فرض جدید از ترجمه بعدی استفاده می‌شود، بدون استقرار مجدد.",
+  "admin.models.savedTitle": "مدل‌ها ذخیره شدند",
+  "admin.models.saving": "در حال ذخیره...",
+  "admin.models.setDefault": "تنظیم {model} به‌عنوان مدل پیش‌فرض",
+  "admin.models.sharedTierNote": "سطح مشترک: اشتراک Business نیز برای موتور «openrouter» از سطح پایه استفاده می‌کند — فهرست بالا برای هر دو اشتراک اعمال می‌شود.",
+  "admin.models.subtitle": "ماتریس plans → سطوح هوش مصنوعی. مدل واقعی استفاده‌شده در هر ترجمه، مدل سطحِ همان اشتراک است: اشتراک Pro هرگز نمی‌تواند مدل Premium را فعال کند.",
+  "admin.models.title": "مدل‌ها و اشتراک‌ها",
+  "admin.stats.aiTiers": "سطوح هوش مصنوعی (۳۰ روز)",
+  "admin.stats.includingCredits": "شامل {amount} اعتبار",
+  "admin.stats.mrr": "درآمد ماهانه تکرارشونده تخمینی",
+  "admin.stats.noTranslationsYet": "هیچ ترجمه‌ای در ۳۰ روز گذشته ثبت نشده است: تفکیک سطوح با ترجمه‌های بعدی کامل می‌شود.",
+  "admin.stats.payments30": "{count} پرداخت در ۳۰ روز",
+  "admin.stats.refreshing": "در حال به‌روزرسانی...",
+  "admin.stats.revenue30": "درآمد (۳۰ روز)",
+  "admin.stats.revenueTotal": "کل درآمد دریافت‌شده",
+  "admin.stats.tiersSub": "پایه / پیشرفته · کلاسیک {classic} · سایر {other}",
+  "admin.stats.unavailable": "آمار کسب‌وکار در دسترس نیست ({error}).",
+  "admin.stats.waitlist": "فهرست انتظار",
+  "admin.stats.waitlistSub": "نفر در انتظار"
 }
--- a/frontend/src/lib/i18n/messages/fa/landing.json
+++ b/frontend/src/lib/i18n/messages/fa/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "برای متخصصان حرفه‌ای",
  "landing.pricing.pro.f1": "۲۰۰ سند / ماه",
  "landing.pricing.pro.f2": "تا ۲۰۰ صفحه برای هر سند",
-  "landing.pricing.pro.f3": "ترجمه مبتنی بر هوش مصنوعی",
+  "landing.pricing.pro.f3": "هوش مصنوعی پایه: DeepSeek V4 Flash، GLM-5.3 Flash، MiniMax M3",
  "landing.pricing.pro.f4": "Google شامل است",
  "landing.pricing.pro.f5": "واژه‌نامه و پرامپت سفارشی",
  "landing.pricing.pro.f6": "پشتیبانی اولویت‌دار",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "برای تیم‌ها با نیاز بالا",
  "landing.pricing.business.f1": "۱,۰۰۰ سند / ماه",
  "landing.pricing.business.f2": "تا ۵۰۰ صفحه برای هر سند",
-  "landing.pricing.business.f3": "هوش مصنوعی پیشرفته (Claude)",
+  "landing.pricing.business.f3": "هوش مصنوعی پیشرفته (Claude 5)",
  "landing.pricing.business.f4": "همه ارائه‌دهندگان + API",
  "landing.pricing.business.f5": "وب‌هوک و خودکارسازی",
  "landing.pricing.business.f6": "۵ صندلی تیمی",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "فایل‌های DOCX, XLSX, PPTX یا PDF پشتیبانی می‌شوند",
  "landing.translate.aiAnalysis": "تحلیل AI فعال",
  "landing.translate.processing": "در حال پردازش",
-  "landing.translate.preservingLayout": "طرح‌بندی شما حفظ می‌شود"
+  "landing.translate.preservingLayout": "طرح‌بندی شما حفظ می‌شود",
+  "landing.beforeAfter.seal": "همان چیدمان، کلمه به کلمه",
+  "landing.beforeAfter.proof1": "نمودارهای SmartArt بازسازی می‌شوند، نه تخت‌شده",
+  "landing.beforeAfter.proof2": "سری‌ها و محورهای نمودار ترجمه می‌شوند",
+  "landing.beforeAfter.proof3": "فهرست مطالب به زبان مقصد بازسازی می‌شود",
+  "landing.beforeAfter.targetTitle": "مشخصات فنی — تهویه هوا",
+  "landing.beforeAfter.targetPara1": "سیستم تهویه باید پیش از پایان سه‌ماهه دوم راه‌اندازی شود.",
+  "landing.beforeAfter.targetPara2Before": "",
+  "landing.beforeAfter.targetTerm": "هواساز مرکزی",
+  "landing.beforeAfter.targetPara2After": "باید مطابق برنامه‌ریزی کلاس فیلتر F7 را رعایت کند.",
+  "landing.beforeAfter.targetItem1": "بار حرارتی: ۴۲ کیلووات در جریان اسمی",
+  "landing.beforeAfter.targetItem2": "سطح صدا کمتر از ۴۵ دسی‌بل (A) در فاصله ۳ متری",
+  "landing.formats.pill": "سازگاری",
+  "landing.hero.visualCaption": "همان چیدمان، زبان جدید — هیچ چیز دیگری تغییر نمی‌کند",
+  "landing.pricing.free.name": "رایگان",
+  "landing.pricing.free.desc": "برای آشنایی با اپلیکیشن ایده‌آل است",
+  "landing.pricing.free.cta": "انتخاب این برنامه",
+  "landing.pricing.enterprise.name": "سازمانی",
+  "landing.pricing.enterprise.desc": "راهکارهای سفارشی برای سازمان‌های بزرگ",
+  "landing.pricing.enterprise.cta": "تماس با ما",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/fa/pricing.json
+++ b/frontend/src/lib/i18n/messages/fa/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "تاریخچه ۹۰ روزه",
  "pricing.plans.business.feat1": "۱,۰۰۰ سند / ماه",
  "pricing.plans.business.feat2": "تا ۵۰۰ صفحه برای هر سند",
-  "pricing.plans.business.feat3": "هوش مصنوعی پایه + پیشرفته (Claude Haiku)",
+  "pricing.plans.business.feat3": "هوش مصنوعی پایه + پیشرفته (Claude 5)",
  "pricing.plans.business.feat4": "تمام ارائه‌دهنده‌های ترجمه",
  "pricing.plans.business.feat5": "فایل‌ها تا ۵۰ مگابایت",
  "pricing.plans.business.feat6": "دسترسی API (۱۰,۰۰۰ فراخوان/ماه)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "تاریخچه ۱ ساله",
  "pricing.plans.business.feat10": "تحلیل پیشرفته",
  "pricing.plans.enterprise.feat1": "اسناد نامحدود",
-  "pricing.plans.enterprise.feat2": "تمام مدل‌های هوش مصنوعی (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "تمام مدل‌های هوش مصنوعی (Claude 5، DeepSeek V4 Pro، GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "استقرار محلی یا ابری اختصاصی",
  "pricing.plans.enterprise.feat4": "SLA 99.9% تضمین‌شده",
  "pricing.plans.enterprise.feat5": "پشتیبانی اختصاصی ۲۴/۷",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "هوش مصنوعی چندنخی فوق‌سریع",
  "pricing.trust.availability.title": "۲۴/۷ در دسترس",
  "pricing.trust.availability.sub": "۹۹.۹٪ زمان فعالیت تضمین‌شده",
-  "pricing.aiModels.title": "مدل‌های هوش مصنوعی ما — مارس ۲۰۲۶",
+  "pricing.aiModels.title": "مدل‌های هوش مصنوعی ما — سپتامبر ۲۰۲۶",
  "pricing.aiModels.essential.title": "ترجمه هوش مصنوعی پایه",
  "pricing.aiModels.essential.plan": "طرح حرفه‌ای",
  "pricing.aiModels.essential.descPrefix": "مبتنی بر",
-  "pricing.aiModels.essential.descSuffix": "— مقرون‌به‌صرفه‌ترین مدل هوش مصنوعی سال ۲۰۲۶. کیفیت قابل مقایسه با مدل‌های پیشرو با کسر هزینه.",
-  "pricing.aiModels.essential.modelName": "مدل هوش مصنوعی Essential",
-  "pricing.aiModels.essential.context": "۱۶۳هزار توکن زمینه",
-  "pricing.aiModels.essential.value": "ارزش عالی نسبت به قیمت",
+  "pricing.aiModels.essential.descSuffix": "— مقرون‌به‌صرفه‌ترین مدل‌های هوش مصنوعی سال ۲۰۲۶.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash، GLM-5.3 Flash و MiniMax M3",
+  "pricing.aiModels.essential.context": "زمینه تا ۱٫۳ میلیون توکن (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "بهترین نسبت کیفیت به قیمت",
  "pricing.aiModels.premium.title": "ترجمه هوش مصنوعی پیشرفته",
  "pricing.aiModels.premium.plan": "طرح سازمانی",
  "pricing.aiModels.premium.descPrefix": "مبتنی بر",
  "pricing.aiModels.premium.descSuffix": "توسط Anthropic — دقیق در اسناد حقوقی، پزشکی و فنی پیچیده.",
-  "pricing.aiModels.premium.context": "۲۰۰هزار توکن زمینه",
+  "pricing.aiModels.premium.context": "یک میلیون توکن زمینه",
  "pricing.aiModels.premium.precision": "بالاترین دقت",
  "pricing.faq.title": "سوالات متداول",
  "pricing.faq.q1": "آیا می‌توانم طرح را هر زمان بخواهم تغییر دهم؟",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "\\\"ترجمه هوش مصنوعی پایه\\\" چیست؟",
  "pricing.faq.a2": "این موتور هوش مصنوعی ماست. زمینه اسناد شما را درک می‌کند، طرح‌بندی را حفظ می‌کند و اصطلاحات فنی را بسیار بهتر از ترجمه کلاسیک مدیریت می‌کند.",
  "pricing.faq.q3": "تفاوت هوش مصنوعی پایه و پیشرفته چیست؟",
-  "pricing.faq.a3": "هوش مصنوعی Essential از یک مدل بهینه‌شده استفاده می‌کند (ارزش عالی برای پول). هوش مصنوعی Premium از Claude 3.5 Haiku انترپیک استفاده می‌کند که در اسناد حقوقی، پزشکی و فنی پیچیده دقیق‌تر است.",
+  "pricing.faq.a3": "هوش مصنوعی Essential از مدل‌های DeepSeek V4 Flash، GLM-5.3 Flash و MiniMax M3 استفاده می‌کند (ارزش عالی برای پول). هوش مصنوعی Premium از Claude 5 انترپیک استفاده می‌کند که در اسناد حقوقی، پزشکی و فنی پیچیده دقیق‌تر است.",
  "pricing.faq.q4": "آیا اسناد من پس از ترجمه نگهداری می‌شوند؟",
  "pricing.faq.a4": "فایل‌های ترجمه‌شده طبق طرح شما در دسترس هستند (۳۰ روز مبتدی، ۹۰ روز حرفه‌ای، ۱ سال سازمانی). آنها در حالت استراحت و هنگام انتقال رمزنگاری می‌شوند.",
  "pricing.faq.q5": "اگر از سهمیه ماهانه خود فراتر رویم چه می‌شود؟",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "خطا در ایجاد پرداخت.",
  "pricing.dashboard": "داشبورد",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "از ۰٫۰۹ دلار به ازای هر میلیون توکن",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "جایگزین‌ها: DeepSeek V4 Pro و GLM-5.3",
+  "pricing.confirm.title": "اشتراک خود را تأیید کنید",
+  "pricing.confirm.subtitle": "به درگاه پرداخت امن ما هدایت می‌شوید.",
+  "pricing.confirm.cancel": "لغو",
+  "pricing.confirm.cta": "ادامه به پرداخت",
+  "pricing.confirm.year": "سال",
+  "pricing.confirm.month": "ماه",
+  "pricing.confirm.monthlyEquivalent": "صورتحساب سالانه — یعنی {price} یورو در ماه.",
+  "pricing.confirm.secureNote": "هر زمان از پروفایل خود قابل لغو است. پرداخت توسط Stripe انجام می‌شود؛ اطلاعات کارت شما هرگز به سرورهای ما نمی‌رسد.",
+  "pricing.enterprise.subject": "پیشنهاد سازمانی",
+  "pricing.enterpriseBand.cta": "تماس با ما",
+  "pricing.enterpriseBand.text": "حجم بالا، موتورهای اختصاصی، گزینه‌های درون‌سازمانی — صحبت کنیم.",
+  "pricing.error.server": "خطای سرور {status}",
+  "pricing.freeBand.cta": "رایگان شروع کنید",
+  "pricing.freeBand.text": "فقط می‌خواهید امتحان کنید؟ رایگان شروع کنید — ماهی ۵ سند، بدون کارت.",
+  "pricing.header.titleBase": "برنامه‌ای برای",
+  "pricing.header.titleAccent": "هر نیاز",
+  "pricing.toast.close": "بستن",
+  "pricing.aiModels.premium.price": "۲ دلار / ۱۰ دلار به ازای هر میلیون توکن"
 }
--- a/frontend/src/lib/i18n/messages/fa/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/fa/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "دسترسی یکپارچه به بهترین مدل‌های متن‌باز بهینه‌سازی‌شده برای ترجمه.",
  "providerTheme.openrouter_premium.badge": "فوق‌العاده",
  "providerTheme.openrouter_premium.subBadge": "حداکثر زمینه",
-  "providerTheme.openrouter_premium.desc": "با کمک مدل‌های پیشرفته (GPT-4o، Claude Sonnet 4.6) برای اسناد طولانی.",
+  "providerTheme.openrouter_premium.desc": "با کمک مدل‌های پیشرفته (GPT-4o، Claude 5) برای اسناد طولانی.",
  "providerTheme.zai.badge": "تخصصی",
  "providerTheme.zai.subBadge": "مالی و حقوقی",
  "providerTheme.zai.desc": "مدل تنظیم‌شده برای اصطلاحات تجاری سخت‌گیرانه (حقوقی، مالی).",
--- a/frontend/src/lib/i18n/messages/fr/admin.json
+++ b/frontend/src/lib/i18n/messages/fr/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Fournisseurs",
  "admin.nav.system": "Système",
  "admin.nav.logs": "Logs",
+  "admin.nav.stats": "Statistiques",
  "admin.users.title": "Gestion des Utilisateurs",
  "admin.users.subtitle": "Visualiser et gérer les comptes utilisateurs",
  "admin.users.planUpdated": "Plan mis à jour",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "En attente de données...",
  "admin.system.purging": "Purge...",
  "admin.system.clean": "Nettoyer",
-  "admin.system.purge": "Purger"
+  "admin.system.purge": "Purger",
+  "admin.nav.models": "Modèles & abonnements",
+  "admin.nav.marketing": "Marketing",
+  "admin.marketing.audience": "Audience",
+  "admin.marketing.audienceAllUsers": "Tous les comptes",
+  "admin.marketing.audienceInactive": "Comptes inactifs depuis 30 jours",
+  "admin.marketing.audiencePlaceholder": "Choisir une audience",
+  "admin.marketing.audiencePlanBusiness": "Forfait Business",
+  "admin.marketing.audiencePlanEnterprise": "Forfait Enterprise",
+  "admin.marketing.audiencePlanFree": "Forfait Gratuit",
+  "admin.marketing.audiencePlanPro": "Forfait Pro",
+  "admin.marketing.audiencePlanStarter": "Forfait Starter",
+  "admin.marketing.audienceWaitlist": "Liste d'attente",
+  "admin.marketing.badgeReal": "réel",
+  "admin.marketing.badgeTest": "test",
+  "admin.marketing.confirmSend": "Envoyer cet email à {count} destinataire(s) (« {audience} ») ?",
+  "admin.marketing.footerNote": "Un pied de page avec le lien de désabonnement est ajouté automatiquement à l'envoi.",
+  "admin.marketing.hidePreview": "Masquer l'aperçu",
+  "admin.marketing.history": "Historique des envois",
+  "admin.marketing.historyFailedSuffix": ", {count} échec(s)",
+  "admin.marketing.historySent": "{sent}/{total} envoyé(s)",
+  "admin.marketing.html": "Contenu HTML",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Renseignez le sujet et le HTML.",
+  "admin.marketing.incompleteTitle": "Contenu incomplet",
+  "admin.marketing.loadErrorTitle": "Erreur de chargement",
+  "admin.marketing.localCheckUnavailable": "Vérification locale indisponible sur ce navigateur : le serveur contrôlera l'envoi test préalable.",
+  "admin.marketing.networkDesc": "Impossible de contacter le backend.",
+  "admin.marketing.newCampaign": "Nouvelle relance",
+  "admin.marketing.noHistory": "Aucun envoi enregistré.",
+  "admin.marketing.preview": "Aperçu",
+  "admin.marketing.previewEmpty": "Aperçu : collez votre HTML ci-contre.",
+  "admin.marketing.previewTitle": "Aperçu de l'email",
+  "admin.marketing.recipientsAvailable": "{count} destinataire(s) disponibles après exclusion des désabonnés.",
+  "admin.marketing.sendReal": "Envoyer à {count} destinataire(s)",
+  "admin.marketing.sendRefusedTitle": "Envoi refusé",
+  "admin.marketing.sendTest": "Envoi test",
+  "admin.marketing.sendingDesc": "{count} email(s) en file d'envoi (0,2 s d'intervalle). Les désabonnés sont exclus.",
+  "admin.marketing.sendingTitle": "Envoi en cours",
+  "admin.marketing.subject": "Sujet",
+  "admin.marketing.subjectPlaceholder": "Ex. : Votre traduction vous attend — 20 % pendant 7 jours",
+  "admin.marketing.subtitle": "Envoi d'emails à une audience, avec envoi test obligatoire, lien de désabonnement automatique et historique complet.",
+  "admin.marketing.testDoneDesc": "Email reçu sur {email}",
+  "admin.marketing.testDoneTitle": "Envoi test effectué",
+  "admin.marketing.testEmail": "Email de test (optionnel)",
+  "admin.marketing.testEmailPlaceholder": "Sinon : adresse d'expédition SMTP",
+  "admin.marketing.testFailedTitle": "Échec de l'envoi test",
+  "admin.marketing.testRequiredHint": "L'envoi réel exige un envoi test préalable de ce contenu exact (vérifié par le serveur).",
+  "admin.marketing.testValidated": "Envoi test validé pour ce contenu.",
+  "admin.marketing.title": "Marketing — relances par email",
+  "admin.marketing.unsubscribedNote": "{count} désabonné(s) seront exclus de tout envoi.",
+  "admin.models.addModel": "Ajouter un modèle",
+  "admin.models.businessBadge": "Forfait Business",
+  "admin.models.cancel": "Annuler",
+  "admin.models.catalog": "Catalogue OpenRouter",
+  "admin.models.catalogErrorDesc": "Impossible de charger le catalogue OpenRouter.",
+  "admin.models.catalogErrorTitle": "Catalogue indisponible",
+  "admin.models.customBadge": "Palier personnalisé",
+  "admin.models.customNote": "Modèles sur mesure, définis au cas par cas avec le client.",
+  "admin.models.defaultBadge": "par défaut",
+  "admin.models.emptyTier": "Aucun modèle : la gamme officielle du plan sera utilisée.",
+  "admin.models.enterpriseBadge": "Forfait Enterprise",
+  "admin.models.essentialBadge": "Palier IA Essentielle",
+  "admin.models.essentialCost": "— facturation coût 1",
+  "admin.models.fallbackFirst": "(secours n°1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — vérifiez votre token admin.",
+  "admin.models.loadErrorTitle": "Erreur de chargement",
+  "admin.models.matrixDesc": "Modèles actifs par palier — l'ordre de la liste est la priorité de secours. Le bouton radio choisit le modèle par défaut, utilisé immédiatement sans redéploiement.",
+  "admin.models.matrixTitle": "Matrice des modèles par forfait",
+  "admin.models.missingConfig": "Configuration des paliers IA introuvable. Rechargez la page.",
+  "admin.models.moveDown": "Descendre",
+  "admin.models.moveUp": "Monter",
+  "admin.models.networkDesc": "Impossible de contacter le backend.",
+  "admin.models.networkTitle": "Erreur réseau",
+  "admin.models.premiumBadge": "Palier IA Premium",
+  "admin.models.premiumCost": "— facturation coût 5",
+  "admin.models.premiumReservedNote": "Réservé aux forfaits Business et Enterprise (moteur « openrouter_premium »).",
+  "admin.models.proBadge": "Forfait Pro",
+  "admin.models.remove": "Retirer",
+  "admin.models.save": "Enregistrer",
+  "admin.models.saveErrorTitle": "Erreur de sauvegarde",
+  "admin.models.saveNetworkDesc": "Impossible d'enregistrer la configuration.",
+  "admin.models.savedDesc": "Le nouveau modèle par défaut est utilisé dès la prochaine traduction, sans redéploiement.",
+  "admin.models.savedTitle": "Modèles enregistrés",
+  "admin.models.saving": "Enregistrement...",
+  "admin.models.setDefault": "Définir {model} comme modèle par défaut",
+  "admin.models.sharedTierNote": "Palier commun : le forfait Business utilise aussi la Essentielle pour son moteur « openrouter » — la liste ci-dessus s'applique aux deux forfaits.",
+  "admin.models.subtitle": "Matrice forfaits → paliers IA. Le modèle réellement utilisé par traduction est celui du palier du plan : un forfait Pro ne peut jamais déclencher un modèle Premium.",
+  "admin.models.title": "Modèles & abonnements",
+  "admin.stats.aiTiers": "Paliers IA (30 j)",
+  "admin.stats.includingCredits": "dont {amount} de crédits",
+  "admin.stats.mrr": "MRR estimé",
+  "admin.stats.noTranslationsYet": "Aucune traduction en base sur les 30 derniers jours : la répartition des paliers se remplira dès les prochaines traductions.",
+  "admin.stats.payments30": "{count} paiement(s) sur 30 jours",
+  "admin.stats.refreshing": "Actualisation...",
+  "admin.stats.revenue30": "Revenus (30 jours)",
+  "admin.stats.revenueTotal": "Revenus encaissés (total)",
+  "admin.stats.tiersSub": "Essentielle / Premium · classique {classic} · autre {other}",
+  "admin.stats.unavailable": "Statistiques business indisponibles ({error}).",
+  "admin.stats.waitlist": "Liste d'attente",
+  "admin.stats.waitlistSub": "inscrits en attente"
 }
--- a/frontend/src/lib/i18n/messages/fr/landing.json
+++ b/frontend/src/lib/i18n/messages/fr/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "Pour les professionnels exigeants",
  "landing.pricing.pro.f1": "200 documents / mois",
  "landing.pricing.pro.f2": "Jusqu'à 200 pages par doc",
-  "landing.pricing.pro.f3": "Traduction par IA",
+  "landing.pricing.pro.f3": "IA Essentielle : DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google inclus",
  "landing.pricing.pro.f5": "Glossaires et prompts",
  "landing.pricing.pro.f6": "Support prioritaire",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "Pour les équipes avec des besoins élevés",
  "landing.pricing.business.f1": "1 000 documents / mois",
  "landing.pricing.business.f2": "Jusqu'à 500 pages par doc",
-  "landing.pricing.business.f3": "IA Premium (Claude)",
+  "landing.pricing.business.f3": "IA Premium (Claude 5)",
  "landing.pricing.business.f4": "Tous les fournisseurs + API",
  "landing.pricing.business.f5": "Webhooks et automatisation",
  "landing.pricing.business.f6": "5 postes d'équipe",
--- a/frontend/src/lib/i18n/messages/fr/pricing.json
+++ b/frontend/src/lib/i18n/messages/fr/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "Historique 90 jours",
  "pricing.plans.business.feat1": "1 000 documents / mois",
  "pricing.plans.business.feat2": "Jusqu'à 500 pages par document",
-  "pricing.plans.business.feat3": "IA Essentielle + Premium (Claude Haiku)",
+  "pricing.plans.business.feat3": "IA Essentielle + IA Premium (Claude 5)",
  "pricing.plans.business.feat4": "Tous les fournisseurs de traduction",
  "pricing.plans.business.feat5": "Fichiers jusqu'à 50 Mo",
  "pricing.plans.business.feat6": "Accès API (10 000 appels/mois)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "Historique 1 an",
  "pricing.plans.business.feat10": "Analytiques avancées",
  "pricing.plans.enterprise.feat1": "Documents illimités",
-  "pricing.plans.enterprise.feat2": "Tous les modèles IA (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "Tous les modèles IA (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "Déploiement on-premise ou cloud dédié",
  "pricing.plans.enterprise.feat4": "SLA 99,9 % garanti",
  "pricing.plans.enterprise.feat5": "Support 24/7 dédié",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "IA multi-thread ultra-rapide",
  "pricing.trust.availability.title": "Disponible 24/7",
  "pricing.trust.availability.sub": "Uptime garanti 99,9 %",
-  "pricing.aiModels.title": "Nos modèles IA — Mars 2026",
+  "pricing.aiModels.title": "Nos modèles IA — Septembre 2026",
  "pricing.aiModels.essential.title": "Traduction IA Essentielle",
  "pricing.aiModels.essential.plan": "Forfait Pro",
  "pricing.aiModels.essential.descPrefix": "Basée sur",
-  "pricing.aiModels.essential.descSuffix": "— le modèle IA le plus rentable de 2026. Qualité comparable aux modèles frontier à une fraction du coût.",
-  "pricing.aiModels.essential.modelName": "notre modèle IA Essentiel",
-  "pricing.aiModels.essential.context": "163K tokens de contexte",
-  "pricing.aiModels.essential.value": "Excellent rapport qualité/prix",
+  "pricing.aiModels.essential.descSuffix": "— les modèles IA au meilleur rapport qualité/prix de 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash et MiniMax M3",
+  "pricing.aiModels.essential.context": "Jusqu'à 1,3 M de contexte (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "Le meilleur rapport qualité/prix",
  "pricing.aiModels.premium.title": "Traduction IA Premium",
  "pricing.aiModels.premium.plan": "Forfait Business",
  "pricing.aiModels.premium.descPrefix": "Basée sur",
  "pricing.aiModels.premium.descSuffix": "d'Anthropic — précis sur les documents juridiques, médicaux et techniques complexes.",
-  "pricing.aiModels.premium.context": "200K tokens de contexte",
+  "pricing.aiModels.premium.context": "1M de contexte",
  "pricing.aiModels.premium.precision": "Meilleure précision",
  "pricing.faq.title": "Questions fréquentes",
  "pricing.faq.q1": "Puis-je changer de forfait à tout moment ?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "Qu'est-ce que la « Traduction IA Essentielle » ?",
  "pricing.faq.a2": "C'est notre moteur IA. Il comprend le contexte de vos documents, préserve la mise en page et gère les termes techniques bien mieux qu'une traduction classique.",
  "pricing.faq.q3": "Quelle est la différence entre IA Essentielle et IA Premium ?",
-  "pricing.faq.a3": "L'IA Essentielle utilise un modèle optimisé (excellent rapport qualité/prix). L'IA Premium utilise Claude Sonnet 4.6 d'Anthropic, plus précis sur les documents juridiques, médicaux et techniques complexes.",
+  "pricing.faq.a3": "L'IA Essentielle s'appuie sur DeepSeek V4 Flash, GLM-5.3 Flash et MiniMax M3 (excellent rapport qualité/prix). L'IA Premium utilise Claude 5 d'Anthropic, plus précis sur les documents juridiques, médicaux et techniques complexes.",
  "pricing.faq.q4": "Mes documents sont-ils conservés après traduction ?",
  "pricing.faq.a4": "Les fichiers traduits sont disponibles selon votre forfait (30 jours Starter, 90 jours Pro, 1 an Business). Ils sont chiffrés au repos et en transit.",
  "pricing.faq.q5": "Que se passe-t-il si je dépasse mon quota mensuel ?",
@@ -163,5 +163,9 @@
  "pricing.freeBand.text": "Juste envie d'essayer ? Commencez gratuitement — 5 documents par mois, sans carte.",
  "pricing.freeBand.cta": "Commencer gratuitement",
  "pricing.enterpriseBand.text": "Volume, moteurs dédiés, hébergement spécifique — parlons-en.",
-  "pricing.enterpriseBand.cta": "Nous contacter"
+  "pricing.enterpriseBand.cta": "Nous contacter",
+  "pricing.aiModels.essential.price": "À partir de 0,09 $ par million de jetons",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Alternatives : DeepSeek V4 Pro et GLM-5.3",
+  "pricing.aiModels.premium.price": "2 $ / 10 $ par 1M de jetons"
 }
--- a/frontend/src/lib/i18n/messages/fr/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/fr/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Accès unifié aux meilleurs modèles open-source optimisés pour la traduction.",
  "providerTheme.openrouter_premium.badge": "Ultra",
  "providerTheme.openrouter_premium.subBadge": "Maximum Context",
-  "providerTheme.openrouter_premium.desc": "Traduction assistée par les modèles de pointe (GPT-4o, Claude Sonnet 4.6) pour documents longs.",
+  "providerTheme.openrouter_premium.desc": "Traduction assistée par les modèles de pointe (GPT-4o, Claude 5) pour documents longs.",
  "providerTheme.zai.badge": "Spécialisée",
  "providerTheme.zai.subBadge": "Finance & Droit",
  "providerTheme.zai.desc": "Modèle affiné pour les terminologies métiers exigeantes (juridique, finance).",
--- a/frontend/src/lib/i18n/messages/it/admin.json
+++ b/frontend/src/lib/i18n/messages/it/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Fornitori",
  "admin.nav.system": "Sistema",
  "admin.nav.logs": "Log",
+  "admin.nav.stats": "Statistiche",
  "admin.users.title": "Gestione Utenti",
  "admin.users.subtitle": "Visualizza e gestisci gli account utente",
  "admin.users.planUpdated": "Piano aggiornato",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "In attesa di dati...",
  "admin.system.purging": "Pulizia in corso...",
  "admin.system.clean": "Pulisci",
-  "admin.system.purge": "Pulisci"
+  "admin.system.purge": "Pulisci",
+  "admin.nav.models": "Modelli e abbonamenti",
+  "admin.nav.marketing": "Marketing",
+  "admin.marketing.audience": "Pubblico",
+  "admin.marketing.audienceAllUsers": "Tutti gli account",
+  "admin.marketing.audienceInactive": "Account inattivi da 30 giorni",
+  "admin.marketing.audiencePlaceholder": "Scegli un pubblico",
+  "admin.marketing.audiencePlanBusiness": "Piano Business",
+  "admin.marketing.audiencePlanEnterprise": "Piano Enterprise",
+  "admin.marketing.audiencePlanFree": "Piano Gratuito",
+  "admin.marketing.audiencePlanPro": "Piano Pro",
+  "admin.marketing.audiencePlanStarter": "Piano Starter",
+  "admin.marketing.audienceWaitlist": "Lista d'attesa",
+  "admin.marketing.badgeReal": "reale",
+  "admin.marketing.badgeTest": "prova",
+  "admin.marketing.confirmSend": "Inviare questa email a {count} destinatario/i («{audience}»)?",
+  "admin.marketing.footerNote": "Un piè di pagina con il link di disiscrizione è aggiunto automaticamente all'invio.",
+  "admin.marketing.hidePreview": "Nascondi anteprima",
+  "admin.marketing.history": "Cronologia degli invii",
+  "admin.marketing.historyFailedSuffix": ", {count} falliti",
+  "admin.marketing.historySent": "{sent}/{total} inviati",
+  "admin.marketing.html": "Contenuto HTML",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Compila l'oggetto e l'HTML.",
+  "admin.marketing.incompleteTitle": "Contenuto incompleto",
+  "admin.marketing.loadErrorTitle": "Errore di caricamento",
+  "admin.marketing.localCheckUnavailable": "Verifica locale non disponibile in questo browser: il server controllerà l'invio di prova previo.",
+  "admin.marketing.networkDesc": "Impossibile contattare il backend.",
+  "admin.marketing.newCampaign": "Nuovo invio",
+  "admin.marketing.noHistory": "Nessun invio registrato.",
+  "admin.marketing.preview": "Anteprima",
+  "admin.marketing.previewEmpty": "Anteprima: incolla qui il tuo HTML.",
+  "admin.marketing.previewTitle": "Anteprima dell'email",
+  "admin.marketing.recipientsAvailable": "{count} destinatario/i disponibili dopo l'esclusione dei disiscritti.",
+  "admin.marketing.sendReal": "Invia a {count} destinatario/i",
+  "admin.marketing.sendRefusedTitle": "Invio rifiutato",
+  "admin.marketing.sendTest": "Invio di prova",
+  "admin.marketing.sendingDesc": "{count} email in coda di invio (intervallo 0,2 s). I disiscritti sono esclusi.",
+  "admin.marketing.sendingTitle": "Invio in corso",
+  "admin.marketing.subject": "Oggetto",
+  "admin.marketing.subjectPlaceholder": "Es.: La tua traduzione ti aspetta — 20% per 7 giorni",
+  "admin.marketing.subtitle": "Invio di email a un pubblico, con invio di prova obbligatorio, link di disiscrizione automatico e cronologia completa.",
+  "admin.marketing.testDoneDesc": "Email ricevuta su {email}",
+  "admin.marketing.testDoneTitle": "Invio di prova effettuato",
+  "admin.marketing.testEmail": "Email di prova (opzionale)",
+  "admin.marketing.testEmailPlaceholder": "Altrimenti: indirizzo mittente SMTP",
+  "admin.marketing.testFailedTitle": "Fallimento dell'invio di prova",
+  "admin.marketing.testRequiredHint": "L'invio reale richiede un invio di prova previo di questo identico contenuto (verificato dal server).",
+  "admin.marketing.testValidated": "Invio di prova convalidato per questo contenuto.",
+  "admin.marketing.title": "Marketing — invii email",
+  "admin.marketing.unsubscribedNote": "{count} contatti disiscritti saranno esclusi da ogni invio.",
+  "admin.models.addModel": "Aggiungi un modello",
+  "admin.models.businessBadge": "Piano Business",
+  "admin.models.cancel": "Annulla",
+  "admin.models.catalog": "Catalogo OpenRouter",
+  "admin.models.catalogErrorDesc": "Impossibile caricare il catalogo OpenRouter.",
+  "admin.models.catalogErrorTitle": "Catalogo non disponibile",
+  "admin.models.customBadge": "Livello personalizzato",
+  "admin.models.customNote": "Modelli su misura, definiti caso per caso con il cliente.",
+  "admin.models.defaultBadge": "predefinito",
+  "admin.models.emptyTier": "Nessun modello: verrà usata la gamma ufficiale del piano.",
+  "admin.models.enterpriseBadge": "Piano Enterprise",
+  "admin.models.essentialBadge": "Livello IA Essenziale",
+  "admin.models.essentialCost": "— fatturato con fattore di costo 1",
+  "admin.models.fallbackFirst": "(fallback n. 1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — verifica il token amministratore.",
+  "admin.models.loadErrorTitle": "Errore di caricamento",
+  "admin.models.matrixDesc": "Modelli attivi per livello — l'ordine della lista è la priorità di fallback. Il pulsante di opzione sceglie il modello predefinito, attivo subito senza ridistribuzione.",
+  "admin.models.matrixTitle": "Matrice dei modelli per piano",
+  "admin.models.missingConfig": "Configurazione dei livelli IA non trovata. Ricarica la pagina.",
+  "admin.models.moveDown": "Sposta giù",
+  "admin.models.moveUp": "Sposta su",
+  "admin.models.networkDesc": "Impossibile contattare il backend.",
+  "admin.models.networkTitle": "Errore di rete",
+  "admin.models.premiumBadge": "Livello IA Premium",
+  "admin.models.premiumCost": "— fatturato con fattore di costo 5",
+  "admin.models.premiumReservedNote": "Riservato ai piani Business ed Enterprise (motore «openrouter_premium»).",
+  "admin.models.proBadge": "Piano Pro",
+  "admin.models.remove": "Rimuovi",
+  "admin.models.save": "Salva",
+  "admin.models.saveErrorTitle": "Errore di salvataggio",
+  "admin.models.saveNetworkDesc": "Impossibile salvare la configurazione.",
+  "admin.models.savedDesc": "Il nuovo modello predefinito è usato dalla prossima traduzione, senza ridistribuzione.",
+  "admin.models.savedTitle": "Modelli salvati",
+  "admin.models.saving": "Salvataggio...",
+  "admin.models.setDefault": "Imposta {model} come modello predefinito",
+  "admin.models.sharedTierNote": "Livello condiviso: il piano Business usa il livello Essenziale anche per il motore «openrouter» — l'elenco sopra vale per entrambi i piani.",
+  "admin.models.subtitle": "Matrice piani → livelli IA. Il modello realmente usato per ogni traduzione è quello del livello del piano: un piano Pro non può mai attivare un modello Premium.",
+  "admin.models.title": "Modelli e abbonamenti",
+  "admin.stats.aiTiers": "Livelli IA (30 g)",
+  "admin.stats.includingCredits": "di cui {amount} di crediti",
+  "admin.stats.mrr": "MRR stimato",
+  "admin.stats.noTranslationsYet": "Nessuna traduzione registrata negli ultimi 30 giorni: la ripartizione per livelli si completerà con le prossime traduzioni.",
+  "admin.stats.payments30": "{count} pagamento/i in 30 giorni",
+  "admin.stats.refreshing": "Aggiornamento...",
+  "admin.stats.revenue30": "Ricavi (30 giorni)",
+  "admin.stats.revenueTotal": "Ricavi incassati (totale)",
+  "admin.stats.tiersSub": "Essenziale / Premium · classico {classic} · altro {other}",
+  "admin.stats.unavailable": "Statistiche business non disponibili ({error}).",
+  "admin.stats.waitlist": "Lista d'attesa",
+  "admin.stats.waitlistSub": "iscritti in attesa"
 }
--- a/frontend/src/lib/i18n/messages/it/landing.json
+++ b/frontend/src/lib/i18n/messages/it/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "Per professionisti esigenti",
  "landing.pricing.pro.f1": "200 documenti / mese",
  "landing.pricing.pro.f2": "Fino a 200 pagine per documento",
-  "landing.pricing.pro.f3": "Traduzione con IA",
+  "landing.pricing.pro.f3": "IA Essenziale: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google inclusi",
  "landing.pricing.pro.f5": "Glossari e prompt personalizzati",
  "landing.pricing.pro.f6": "Supporto prioritario",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "Per team con esigenze elevate",
  "landing.pricing.business.f1": "1 000 documenti / mese",
  "landing.pricing.business.f2": "Fino a 500 pagine per documento",
-  "landing.pricing.business.f3": "IA Premium (Claude)",
+  "landing.pricing.business.f3": "IA Premium (Claude 5)",
  "landing.pricing.business.f4": "Tutti i provider + accesso API",
  "landing.pricing.business.f5": "Webhook e automazione",
  "landing.pricing.business.f6": "5 postazioni team",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "File DOCX, XLSX, PPTX o PDF supportati",
  "landing.translate.aiAnalysis": "Analisi IA Attiva",
  "landing.translate.processing": "Elaborazione in corso",
-  "landing.translate.preservingLayout": "Il tuo layout viene preservato"
+  "landing.translate.preservingLayout": "Il tuo layout viene preservato",
+  "landing.beforeAfter.seal": "Stesso layout, parola per parola",
+  "landing.beforeAfter.proof1": "I diagrammi SmartArt vengono ricostruiti, non appiattiti",
+  "landing.beforeAfter.proof2": "Serie e assi dei grafici tradotti",
+  "landing.beforeAfter.proof3": "Indici rigenerati nella lingua di destinazione",
+  "landing.beforeAfter.targetTitle": "Specifiche tecniche — Trattamento aria",
+  "landing.beforeAfter.targetPara1": "L'impianto di ventilazione deve essere messo in servizio entro la fine del secondo trimestre.",
+  "landing.beforeAfter.targetPara2Before": "",
+  "landing.beforeAfter.targetTerm": "L'unità di trattamento aria",
+  "landing.beforeAfter.targetPara2After": "deve rispettare la classe di filtraggio F7 come da programma.",
+  "landing.beforeAfter.targetItem1": "Carico termico: 42 kW alla portata nominale",
+  "landing.beforeAfter.targetItem2": "Livello sonoro inferiore a 45 dB(A) a 3 m",
+  "landing.formats.pill": "COMPATIBILITÀ",
+  "landing.hero.visualCaption": "Stesso layout, nuova lingua — il resto non si tocca",
+  "landing.pricing.free.name": "Gratuito",
+  "landing.pricing.free.desc": "Ideale per scoprire l'app",
+  "landing.pricing.free.cta": "Scegli questo piano",
+  "landing.pricing.enterprise.name": "Enterprise",
+  "landing.pricing.enterprise.desc": "Soluzioni su misura per grandi organizzazioni",
+  "landing.pricing.enterprise.cta": "Contattaci",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/it/pricing.json
+++ b/frontend/src/lib/i18n/messages/it/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "Cronologia di 90 giorni",
  "pricing.plans.business.feat1": "1 000 documenti / mese",
  "pricing.plans.business.feat2": "Fino a 500 pagine per documento",
-  "pricing.plans.business.feat3": "IA Essenziale + Premium (Claude Haiku)",
+  "pricing.plans.business.feat3": "IA Essenziale + IA Premium (Claude 5)",
  "pricing.plans.business.feat4": "Tutti i provider di traduzione",
  "pricing.plans.business.feat5": "File fino a 50 MB",
  "pricing.plans.business.feat6": "Accesso API (10 000 chiamate/mese)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "Cronologia di 1 anno",
  "pricing.plans.business.feat10": "Analisi avanzate",
  "pricing.plans.enterprise.feat1": "Documenti illimitati",
-  "pricing.plans.enterprise.feat2": "Tutti i modelli IA (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "Tutti i modelli IA (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "Distribuzione on-premise o cloud dedicato",
  "pricing.plans.enterprise.feat4": "SLA 99,9 % garantito",
  "pricing.plans.enterprise.feat5": "Supporto dedicato 24/7",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "IA multi-thread ultraveloce",
  "pricing.trust.availability.title": "Disponibile 24/7",
  "pricing.trust.availability.sub": "99,9 % di uptime garantito",
-  "pricing.aiModels.title": "I nostri modelli IA — Marzo 2026",
+  "pricing.aiModels.title": "I nostri modelli IA — Settembre 2026",
  "pricing.aiModels.essential.title": "Traduzione IA Essenziale",
  "pricing.aiModels.essential.plan": "Piano Pro",
  "pricing.aiModels.essential.descPrefix": "Basato su",
-  "pricing.aiModels.essential.descSuffix": "— il modello IA più conveniente del 2026. Qualità paragonabile ai modelli frontier a una frazione del costo.",
-  "pricing.aiModels.essential.modelName": "il nostro modello IA Essenziale",
-  "pricing.aiModels.essential.context": "163K token di contesto",
-  "pricing.aiModels.essential.value": "Eccellente rapporto qualità-prezzo",
+  "pricing.aiModels.essential.descSuffix": "— i modelli IA con il miglior rapporto qualità-prezzo del 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash e MiniMax M3",
+  "pricing.aiModels.essential.context": "Fino a 1,3 M di contesto (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "Il miglior rapporto qualità-prezzo",
  "pricing.aiModels.premium.title": "Traduzione IA Premium",
  "pricing.aiModels.premium.plan": "Piano Business",
  "pricing.aiModels.premium.descPrefix": "Basato su",
  "pricing.aiModels.premium.descSuffix": "di Anthropic — preciso su documenti legali, medici e tecnici complessi.",
-  "pricing.aiModels.premium.context": "200K token di contesto",
+  "pricing.aiModels.premium.context": "1M di contesto",
  "pricing.aiModels.premium.precision": "Massima precisione",
  "pricing.faq.title": "Domande frequenti",
  "pricing.faq.q1": "Posso cambiare piano in qualsiasi momento?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "Cos'è la «Traduzione IA Essenziale»?",
  "pricing.faq.a2": "È il nostro motore IA. Comprende il contesto dei vostri documenti, preserva il layout e gestisce i termini tecnici molto meglio della traduzione classica.",
  "pricing.faq.q3": "Qual è la differenza tra IA Essenziale e IA Premium?",
-  "pricing.faq.a3": "La IA Essenziale usa un modello ottimizzato (eccellente rapporto qualità/prezzo). La IA Premium usa Claude 3.5 Haiku di Anthropic, più precisa su documenti legali, medici e tecnici complessi.",
+  "pricing.faq.a3": "La IA Essenziale si basa su DeepSeek V4 Flash, GLM-5.3 Flash e MiniMax M3 (eccellente rapporto qualità/prezzo). La IA Premium usa Claude 5 di Anthropic, più precisa su documenti legali, medici e tecnici complessi.",
  "pricing.faq.q4": "I miei documenti vengono conservati dopo la traduzione?",
  "pricing.faq.a4": "I file tradotti sono disponibili secondo il tuo piano (30 giorni Starter, 90 giorni Pro, 1 anno Business). Sono crittografati a riposo e in transito.",
  "pricing.faq.q5": "Cosa succede se supero la quota mensile?",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "Errore durante la creazione del pagamento.",
  "pricing.dashboard": "Dashboard",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "Da 0,09 $ per milione di token",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Alternative: DeepSeek V4 Pro e GLM-5.3",
+  "pricing.confirm.title": "Conferma il tuo abbonamento",
+  "pricing.confirm.subtitle": "Verrai reindirizzato al nostro fornitore di pagamento sicuro.",
+  "pricing.confirm.cancel": "Annulla",
+  "pricing.confirm.cta": "Vai al pagamento",
+  "pricing.confirm.year": "anno",
+  "pricing.confirm.month": "mese",
+  "pricing.confirm.monthlyEquivalent": "Fatturazione annuale — ovvero {price} € / mese.",
+  "pricing.confirm.secureNote": "Disdicibile in qualsiasi momento dal tuo profilo. Il pagamento è elaborato da Stripe; i dati della tua carta non passano mai dai nostri server.",
+  "pricing.enterprise.subject": "Richiesta Enterprise",
+  "pricing.enterpriseBand.cta": "Contattaci",
+  "pricing.enterpriseBand.text": "Volume, motori dedicati, opzioni on-premise — parliamone.",
+  "pricing.error.server": "Errore del server {status}",
+  "pricing.freeBand.cta": "Inizia gratis",
+  "pricing.freeBand.text": "Vuoi solo provare? Inizia gratis — 5 documenti al mese, senza carta.",
+  "pricing.header.titleBase": "Un piano per",
+  "pricing.header.titleAccent": "ogni esigenza",
+  "pricing.toast.close": "Chiudi",
+  "pricing.aiModels.premium.price": "2 $ / 10 $ per 1M di token"
 }
--- a/frontend/src/lib/i18n/messages/it/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/it/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Accesso unificato ai migliori modelli open-source ottimizzati per la traduzione.",
  "providerTheme.openrouter_premium.badge": "Ultra",
  "providerTheme.openrouter_premium.subBadge": "Contesto massimo",
-  "providerTheme.openrouter_premium.desc": "Assistito da modelli all'avanguardia (GPT-4o, Claude Sonnet 4.6) per documenti lunghi.",
+  "providerTheme.openrouter_premium.desc": "Assistito da modelli all'avanguardia (GPT-4o, Claude 5) per documenti lunghi.",
  "providerTheme.zai.badge": "Specializzata",
  "providerTheme.zai.subBadge": "Finanza e Diritto",
  "providerTheme.zai.desc": "Modello ottimizzato per terminologie aziendali impegnative (legale, finanza).",
--- a/frontend/src/lib/i18n/messages/ja/admin.json
+++ b/frontend/src/lib/i18n/messages/ja/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "プロバイダー",
  "admin.nav.system": "システム",
  "admin.nav.logs": "ログ",
+  "admin.nav.stats": "統計",
  "admin.users.title": "ユーザー管理",
  "admin.users.subtitle": "ユーザーアカウントの表示と管理",
  "admin.users.planUpdated": "プランが更新されました",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "データを待機中...",
  "admin.system.purging": "パージ中...",
  "admin.system.clean": "クリーン",
-  "admin.system.purge": "パージ"
+  "admin.system.purge": "パージ",
+  "admin.nav.models": "モデル & サブスクリプション",
+  "admin.nav.marketing": "マーケティング",
+  "admin.marketing.audience": "対象者",
+  "admin.marketing.audienceAllUsers": "全アカウント",
+  "admin.marketing.audienceInactive": "30日間非アクティブなアカウント",
+  "admin.marketing.audiencePlaceholder": "対象者を選択",
+  "admin.marketing.audiencePlanBusiness": "Businessプラン",
+  "admin.marketing.audiencePlanEnterprise": "Enterpriseプラン",
+  "admin.marketing.audiencePlanFree": "無料プラン",
+  "admin.marketing.audiencePlanPro": "Proプラン",
+  "admin.marketing.audiencePlanStarter": "Starterプラン",
+  "admin.marketing.audienceWaitlist": "ウェイトリスト",
+  "admin.marketing.badgeReal": "本番",
+  "admin.marketing.badgeTest": "テスト",
+  "admin.marketing.confirmSend": "このメールを {count} 名（「{audience}」）に送信しますか？",
+  "admin.marketing.footerNote": "送信時、配信停止リンク付きフッターが自動で追加されます。",
+  "admin.marketing.hidePreview": "プレビューを隠す",
+  "admin.marketing.history": "送信履歴",
+  "admin.marketing.historyFailedSuffix": "、{count} 件失敗",
+  "admin.marketing.historySent": "{sent}/{total} 送信済み",
+  "admin.marketing.html": "HTMLコンテンツ",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "件名とHTMLを入力してください。",
+  "admin.marketing.incompleteTitle": "内容が不完全です",
+  "admin.marketing.loadErrorTitle": "読み込みエラー",
+  "admin.marketing.localCheckUnavailable": "このブラウザではローカル確認ができないため、サーバーが事前テスト送信を検証します。",
+  "admin.marketing.networkDesc": "バックエンドに接続できません。",
+  "admin.marketing.newCampaign": "新しいキャンペーン",
+  "admin.marketing.noHistory": "送信履歴はありません。",
+  "admin.marketing.preview": "プレビュー",
+  "admin.marketing.previewEmpty": "プレビュー：左側にHTMLを貼り付けてください。",
+  "admin.marketing.previewTitle": "メールプレビュー",
+  "admin.marketing.recipientsAvailable": "配信停止を除いた {count} 名に送信できます。",
+  "admin.marketing.sendReal": "{count} 名に送信",
+  "admin.marketing.sendRefusedTitle": "送信が拒否されました",
+  "admin.marketing.sendTest": "テスト送信",
+  "admin.marketing.sendingDesc": "{count} 件が送信キューに入りました（0.2秒間隔）。配信停止中の宛先は除外されます。",
+  "admin.marketing.sendingTitle": "送信中",
+  "admin.marketing.subject": "件名",
+  "admin.marketing.subjectPlaceholder": "例：翻訳ができあがっています — 7日間20%オフ",
+  "admin.marketing.subtitle": "対象者へのメール送信。テスト送信が必須で、配信停止リンク自動付加、完全な履歴付き。",
+  "admin.marketing.testDoneDesc": "{email} でメールを受信",
+  "admin.marketing.testDoneTitle": "テスト送信が完了しました",
+  "admin.marketing.testEmail": "テストメール（任意）",
+  "admin.marketing.testEmailPlaceholder": "未指定の場合：SMTP送信者アドレス",
+  "admin.marketing.testFailedTitle": "テスト送信に失敗しました",
+  "admin.marketing.testRequiredHint": "本送信には、同一内容の事前テスト送信が必要です（サーバーが検証します）。",
+  "admin.marketing.testValidated": "この内容のテスト送信は検証済みです。",
+  "admin.marketing.title": "マーケティング — メールキャンペーン",
+  "admin.marketing.unsubscribedNote": "配信停止中の {count} 件はすべての送信から除外されます。",
+  "admin.models.addModel": "モデルを追加",
+  "admin.models.businessBadge": "Businessプラン",
+  "admin.models.cancel": "キャンセル",
+  "admin.models.catalog": "OpenRouterカタログ",
+  "admin.models.catalogErrorDesc": "OpenRouterカタログを読み込めませんでした。",
+  "admin.models.catalogErrorTitle": "カタログを利用できません",
+  "admin.models.customBadge": "カスタムティア",
+  "admin.models.customNote": "カスタムモデルはお客様ごとに個別に決定します。",
+  "admin.models.defaultBadge": "既定",
+  "admin.models.emptyTier": "モデルなし：プランの公式ラインナップが使用されます。",
+  "admin.models.enterpriseBadge": "Enterpriseプラン",
+  "admin.models.essentialBadge": "エッセンシャルAIティア",
+  "admin.models.essentialCost": "— コスト係数 1 で課金",
+  "admin.models.fallbackFirst": "（フォールバック第1位）",
+  "admin.models.loadErrorDesc": "HTTP {status} — 管理者トークンを確認してください。",
+  "admin.models.loadErrorTitle": "読み込みエラー",
+  "admin.models.matrixDesc": "ティアごとの有効モデル — リストの順序がフォールバック優先度です。ラジオボタンで既定モデルを選ぶと、再デプロイなしですぐに反映されます。",
+  "admin.models.matrixTitle": "プラン別モデルマトリクス",
+  "admin.models.missingConfig": "AIティアの設定が見つかりません。ページを再読み込みしてください。",
+  "admin.models.moveDown": "下へ移動",
+  "admin.models.moveUp": "上へ移動",
+  "admin.models.networkDesc": "バックエンドに接続できません。",
+  "admin.models.networkTitle": "ネットワークエラー",
+  "admin.models.premiumBadge": "プレミアムAIティア",
+  "admin.models.premiumCost": "— コスト係数 5 で課金",
+  "admin.models.premiumReservedNote": "Business・Enterpriseプラン専用（「openrouter_premium」エンジン）。",
+  "admin.models.proBadge": "Proプラン",
+  "admin.models.remove": "削除",
+  "admin.models.save": "保存",
+  "admin.models.saveErrorTitle": "保存エラー",
+  "admin.models.saveNetworkDesc": "設定を保存できませんでした。",
+  "admin.models.savedDesc": "新しい既定モデルは次の翻訳から使用されます（再デプロイ不要）。",
+  "admin.models.savedTitle": "モデルを保存しました",
+  "admin.models.saving": "保存中...",
+  "admin.models.setDefault": "{model} を既定モデルに設定",
+  "admin.models.sharedTierNote": "共通ティア：Businessプランも「openrouter」エンジンにはエッセンシャルティアを使います — 上のリストは両プランに適用されます。",
+  "admin.models.subtitle": "プラン → AIティアのマトリクス。翻訳ごとに実際に使われるモデルはプランのティアのモデルです：Proプランがプレミアムモデルを呼び出すことは決してありません。",
+  "admin.models.title": "モデル & サブスクリプション",
+  "admin.stats.aiTiers": "AIティア（30日）",
+  "admin.stats.includingCredits": "うちクレジット {amount}",
+  "admin.stats.mrr": "推定MRR",
+  "admin.stats.noTranslationsYet": "過去30日間の翻訳記録がありません：今後の翻訳でティア別内訳が埋まります。",
+  "admin.stats.payments30": "30日間の支払い {count} 件",
+  "admin.stats.refreshing": "更新中...",
+  "admin.stats.revenue30": "売上（30日）",
+  "admin.stats.revenueTotal": "累計売上",
+  "admin.stats.tiersSub": "エッセンシャル / プレミアム · クラシック {classic} · その他 {other}",
+  "admin.stats.unavailable": "ビジネス統計を利用できません（{error}）。",
+  "admin.stats.waitlist": "ウェイトリスト",
+  "admin.stats.waitlistSub": "待機中の登録者"
 }
--- a/frontend/src/lib/i18n/messages/ja/landing.json
+++ b/frontend/src/lib/i18n/messages/ja/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "プロフェッショナル向け",
  "landing.pricing.pro.f1": "月200ドキュメント",
  "landing.pricing.pro.f2": "1ドキュメント最大200ページ",
-  "landing.pricing.pro.f3": "AI翻訳",
+  "landing.pricing.pro.f3": "エッセンシャルAI：DeepSeek V4 Flash、GLM-5.3 Flash、MiniMax M3",
  "landing.pricing.pro.f4": "Google込み",
  "landing.pricing.pro.f5": "カスタム用語集とプロンプト",
  "landing.pricing.pro.f6": "優先サポート",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "大量処理が必要なチーム向け",
  "landing.pricing.business.f1": "月1,000ドキュメント",
  "landing.pricing.business.f2": "1ドキュメント最大500ページ",
-  "landing.pricing.business.f3": "プレミアムAI（Claude）",
+  "landing.pricing.business.f3": "プレミアムAI（Claude 5）",
  "landing.pricing.business.f4": "全プロバイダー + API",
  "landing.pricing.business.f5": "Webhookと自動化",
  "landing.pricing.business.f6": "チームシート5席",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "DOCX, XLSX, PPTX, PDFファイル対応",
  "landing.translate.aiAnalysis": "AI分析アクティブ",
  "landing.translate.processing": "処理中",
-  "landing.translate.preservingLayout": "レイアウトを保持しています"
+  "landing.translate.preservingLayout": "レイアウトを保持しています",
+  "landing.beforeAfter.seal": "同じレイアウト、一語一句そのまま",
+  "landing.beforeAfter.proof1": "SmartArt図面は平面化されず、再構築されます",
+  "landing.beforeAfter.proof2": "グラフの系列と軸も翻訳",
+  "landing.beforeAfter.proof3": "目次は翻訳先言語で再生成",
+  "landing.beforeAfter.targetTitle": "技術仕様 — 空調処理",
+  "landing.beforeAfter.targetPara1": "換気設備は第2四半期末までに稼働させなければなりません。",
+  "landing.beforeAfter.targetPara2Before": "",
+  "landing.beforeAfter.targetTerm": "空調機",
+  "landing.beforeAfter.targetPara2After": "は計画どおりフィルタクラス F7 を満たす必要があります。",
+  "landing.beforeAfter.targetItem1": "熱負荷：定格流量で42 kW",
+  "landing.beforeAfter.targetItem2": "騒音レベル：3 mで45 dB(A)未満",
+  "landing.formats.pill": "対応フォーマット",
+  "landing.hero.visualCaption": "同じレイアウト、新しい言語 — 他は一切変わりません",
+  "landing.pricing.free.name": "無料",
+  "landing.pricing.free.desc": "アプリのお試しに最適",
+  "landing.pricing.free.cta": "このプランを選択",
+  "landing.pricing.enterprise.name": "エンタープライズ",
+  "landing.pricing.enterprise.desc": "大規模組織向けのカスタムソリューション",
+  "landing.pricing.enterprise.cta": "お問い合わせ",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/ja/pricing.json
+++ b/frontend/src/lib/i18n/messages/ja/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "90日間の履歴",
  "pricing.plans.business.feat1": "月1,000ドキュメント",
  "pricing.plans.business.feat2": "1ドキュメントあたり最大500ページ",
-  "pricing.plans.business.feat3": "エッセンシャル + プレミアムAI（Claude Haiku）",
+  "pricing.plans.business.feat3": "エッセンシャル + プレミアムAI（Claude 5）",
  "pricing.plans.business.feat4": "全翻訳プロバイダー",
  "pricing.plans.business.feat5": "最大50 MBのファイル",
  "pricing.plans.business.feat6": "APIアクセス（月10,000コール）",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "1年間の履歴",
  "pricing.plans.business.feat10": "高度な分析",
  "pricing.plans.enterprise.feat1": "無制限ドキュメント",
-  "pricing.plans.enterprise.feat2": "全AIモデル（GPT-5、Claude Opus 4.6…）",
+  "pricing.plans.enterprise.feat2": "全AIモデル（Claude 5、DeepSeek V4 Pro、GLM-5.3…）",
  "pricing.plans.enterprise.feat3": "オンプレミスまたは専用クラウド展開",
  "pricing.plans.enterprise.feat4": "99.9% SLA保証",
  "pricing.plans.enterprise.feat5": "24/7専任サポート",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "超高速マルチスレッドAI",
  "pricing.trust.availability.title": "24/7利用可能",
  "pricing.trust.availability.sub": "99.9%稼働率保証",
-  "pricing.aiModels.title": "AIモデル一覧 — 2026年3月",
+  "pricing.aiModels.title": "AIモデル一覧 — 2026年9月",
  "pricing.aiModels.essential.title": "エッセンシャルAI翻訳",
  "pricing.aiModels.essential.plan": "Proプラン",
  "pricing.aiModels.essential.descPrefix": "ベース：",
-  "pricing.aiModels.essential.descSuffix": "— 2026年で最も費用対効果の高いAIモデル。フロンティアモデルと同等の品質をわずかなコストで実現。",
-  "pricing.aiModels.essential.modelName": "Essential AIモデル",
-  "pricing.aiModels.essential.context": "163Kトークンのコンテキスト",
-  "pricing.aiModels.essential.value": "優れた費用対効果",
+  "pricing.aiModels.essential.descSuffix": "— 2026年で最も費用対効果の高いAIモデル群。",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash、GLM-5.3 Flash、MiniMax M3",
+  "pricing.aiModels.essential.context": "最大130万トークンのコンテキスト（GLM-5.3 Flash）",
+  "pricing.aiModels.essential.value": "最高のコストパフォーマンス",
  "pricing.aiModels.premium.title": "プレミアムAI翻訳",
  "pricing.aiModels.premium.plan": "Businessプラン",
  "pricing.aiModels.premium.descPrefix": "ベース：",
  "pricing.aiModels.premium.descSuffix": "（Anthropic社製）— 法務、医療、複雑な技術文書に高精度。",
-  "pricing.aiModels.premium.context": "200Kトークンのコンテキスト",
+  "pricing.aiModels.premium.context": "100万トークンのコンテキスト",
  "pricing.aiModels.premium.precision": "最高精度",
  "pricing.faq.title": "よくある質問",
  "pricing.faq.q1": "いつでもプランを変更できますか？",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "「エッセンシャルAI翻訳」とは何ですか？",
  "pricing.faq.a2": "独自のAIエンジンです。ドキュメントの文脈を理解し、レイアウトを保持し、技術用語を従来の翻訳よりはるかに適切に処理します。",
  "pricing.faq.q3": "エッセンシャルAIとプレミアムAIの違いは何ですか？",
-  "pricing.faq.a3": "Essential AIは最適化されたモデルを使用しています（コストパフォーマンスに優れています）。Premium AIはAnthropicのClaude 3.5 Haikuを使用し、法的、医学的、複雑な技術文書により正確です。",
+  "pricing.faq.a3": "Essential AIはDeepSeek V4 Flash、GLM-5.3 Flash、MiniMax M3を使用しています（コストパフォーマンスに優れています）。Premium AIはAnthropicのClaude 5を使用し、法的、医学的、複雑な技術文書により正確です。",
  "pricing.faq.q4": "翻訳後もドキュメントは保存されますか？",
  "pricing.faq.a4": "翻訳済みファイルはプランに応じて利用可能です（Starter 30日、Pro 90日、Business 1年）。保存時および通信時は暗号化されています。",
  "pricing.faq.q5": "月間枠を超えた場合はどうなりますか？",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "支払いの作成中にエラーが発生しました。",
  "pricing.dashboard": "ダッシュボード",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "100万トークンあたり$0.09から",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "代替モデル：DeepSeek V4 Pro、GLM-5.3",
+  "pricing.confirm.title": "サブスクリプションの確認",
+  "pricing.confirm.subtitle": "安全な決済プロバイダーへ移動します。",
+  "pricing.confirm.cancel": "キャンセル",
+  "pricing.confirm.cta": "支払いへ進む",
+  "pricing.confirm.year": "年",
+  "pricing.confirm.month": "月",
+  "pricing.confirm.monthlyEquivalent": "年間請求 — 月あたり {price} ユーロ相当。",
+  "pricing.confirm.secureNote": "プロフィールからいつでも解約できます。決済はStripeが処理し、カード情報が当社のサーバーに送信されることはありません。",
+  "pricing.enterprise.subject": "エンタープライズプランについてのお問い合わせ",
+  "pricing.enterpriseBand.cta": "お問い合わせ",
+  "pricing.enterpriseBand.text": "大量処理、専用エンジン、オンプレミス対応 — ご相談ください。",
+  "pricing.error.server": "サーバーエラー {status}",
+  "pricing.freeBand.cta": "無料で始める",
+  "pricing.freeBand.text": "まず試してみたい方へ — 毎月5件まで無料、カード不要。",
+  "pricing.header.titleBase": "あらゆるニーズに",
+  "pricing.header.titleAccent": "ぴったりのプラン",
+  "pricing.toast.close": "閉じる",
+  "pricing.aiModels.premium.price": "100万トークンあたり $2 / $10"
 }
--- a/frontend/src/lib/i18n/messages/ja/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/ja/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "翻訳に最適化された最高のオープンソースモデルへの統一アクセス。",
  "providerTheme.openrouter_premium.badge": "ウルトラ",
  "providerTheme.openrouter_premium.subBadge": "最大コンテキスト",
-  "providerTheme.openrouter_premium.desc": "GPT-4o、Claude Sonnet 4.6 などの最先端モデルによる長文支援。",
+  "providerTheme.openrouter_premium.desc": "GPT-4o、Claude 5 などの最先端モデルによる長文支援。",
  "providerTheme.zai.badge": "特化型",
  "providerTheme.zai.subBadge": "金融と法律",
  "providerTheme.zai.desc": "要求の厳しいビジネス用語 (法務、金融) 向けに微調整されたモデル。",
--- a/frontend/src/lib/i18n/messages/ko/admin.json
+++ b/frontend/src/lib/i18n/messages/ko/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "제공업체",
  "admin.nav.system": "시스템",
  "admin.nav.logs": "로그",
+  "admin.nav.stats": "통계",
  "admin.users.title": "사용자 관리",
  "admin.users.subtitle": "사용자 계정 보기 및 관리",
  "admin.users.planUpdated": "플랜이 업데이트되었습니다",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "데이터 대기 중...",
  "admin.system.purging": "삭제 중...",
  "admin.system.clean": "정리",
-  "admin.system.purge": "삭제"
+  "admin.system.purge": "삭제",
+  "admin.nav.models": "모델 & 구독",
+  "admin.nav.marketing": "마케팅",
+  "admin.marketing.audience": "대상",
+  "admin.marketing.audienceAllUsers": "전체 계정",
+  "admin.marketing.audienceInactive": "30일간 비활성 계정",
+  "admin.marketing.audiencePlaceholder": "대상 선택",
+  "admin.marketing.audiencePlanBusiness": "Business 플랜",
+  "admin.marketing.audiencePlanEnterprise": "Enterprise 플랜",
+  "admin.marketing.audiencePlanFree": "무료 플랜",
+  "admin.marketing.audiencePlanPro": "Pro 플랜",
+  "admin.marketing.audiencePlanStarter": "Starter 플랜",
+  "admin.marketing.audienceWaitlist": "대기 명단",
+  "admin.marketing.badgeReal": "실제",
+  "admin.marketing.badgeTest": "테스트",
+  "admin.marketing.confirmSend": "이 메일을 {count}명(«{audience}»)에게 보내시겠습니까?",
+  "admin.marketing.footerNote": "발송 시 수신거부 링크가 있는 바닥글이 자동으로 추가됩니다.",
+  "admin.marketing.hidePreview": "미리보기 숨기기",
+  "admin.marketing.history": "발송 이력",
+  "admin.marketing.historyFailedSuffix": ", {count}건 실패",
+  "admin.marketing.historySent": "{sent}/{total} 발송",
+  "admin.marketing.html": "HTML 내용",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "제목과 HTML을 입력하세요.",
+  "admin.marketing.incompleteTitle": "내용 불완전",
+  "admin.marketing.loadErrorTitle": "불러오기 오류",
+  "admin.marketing.localCheckUnavailable": "이 브라우저에서는 로컬 확인이 불가하여 서버가 사전 테스트 발송을 검증합니다.",
+  "admin.marketing.networkDesc": "백엔드에 연결할 수 없습니다.",
+  "admin.marketing.newCampaign": "새 캠페인",
+  "admin.marketing.noHistory": "발송 기록이 없습니다.",
+  "admin.marketing.preview": "미리보기",
+  "admin.marketing.previewEmpty": "미리보기: 왼쪽에 HTML을 붙여넣으세요.",
+  "admin.marketing.previewTitle": "이메일 미리보기",
+  "admin.marketing.recipientsAvailable": "수신거부 제외 후 {count}명에게 발송할 수 있습니다.",
+  "admin.marketing.sendReal": "{count}명에게 보내기",
+  "admin.marketing.sendRefusedTitle": "발송이 거부되었습니다",
+  "admin.marketing.sendTest": "테스트 발송",
+  "admin.marketing.sendingDesc": "{count}건이 발송 대기열에 있습니다(0.2초 간격). 수신거부는 제외됩니다.",
+  "admin.marketing.sendingTitle": "발송 진행 중",
+  "admin.marketing.subject": "제목",
+  "admin.marketing.subjectPlaceholder": "예: 번역이 기다리고 있습니다 — 7일간 20% 할인",
+  "admin.marketing.subtitle": "대상에게 이메일을 보냅니다. 필수 테스트 발송, 자동 수신거부 링크, 전체 이력이 포함됩니다.",
+  "admin.marketing.testDoneDesc": "{email}에서 메일 수신",
+  "admin.marketing.testDoneTitle": "테스트 발송 완료",
+  "admin.marketing.testEmail": "테스트 이메일 (선택)",
+  "admin.marketing.testEmailPlaceholder": "미입력 시: SMTP 발신 주소",
+  "admin.marketing.testFailedTitle": "테스트 발송 실패",
+  "admin.marketing.testRequiredHint": "실제 발송에는 동일 내용의 사전 테스트 발송이 필요합니다(서버가 검증).",
+  "admin.marketing.testValidated": "이 내용의 테스트 발송이 검증되었습니다.",
+  "admin.marketing.title": "마케팅 — 이메일 캠페인",
+  "admin.marketing.unsubscribedNote": "수신거부 {count}건은 모든 발송에서 제외됩니다.",
+  "admin.models.addModel": "모델 추가",
+  "admin.models.businessBadge": "Business 플랜",
+  "admin.models.cancel": "취소",
+  "admin.models.catalog": "OpenRouter 카탈로그",
+  "admin.models.catalogErrorDesc": "OpenRouter 카탈로그를 불러오지 못했습니다.",
+  "admin.models.catalogErrorTitle": "카탈로그 사용 불가",
+  "admin.models.customBadge": "맞춤 등급",
+  "admin.models.customNote": "맞춤 모델은 고객과 사례별로 정의합니다.",
+  "admin.models.defaultBadge": "기본",
+  "admin.models.emptyTier": "모델 없음: 플랜의 공식 라인업이 사용됩니다.",
+  "admin.models.enterpriseBadge": "Enterprise 플랜",
+  "admin.models.essentialBadge": "에센셜 AI 등급",
+  "admin.models.essentialCost": "— 비용 계수 1로 청구",
+  "admin.models.fallbackFirst": "(폴백 1순위)",
+  "admin.models.loadErrorDesc": "HTTP {status} — 관리자 토큰을 확인하세요.",
+  "admin.models.loadErrorTitle": "불러오기 오류",
+  "admin.models.matrixDesc": "등급별 활성 모델 — 목록 순서가 폴백 우선순위입니다. 라디오 버튼으로 기본 모델을 선택하면 재배포 없이 즉시 적용됩니다.",
+  "admin.models.matrixTitle": "플랜별 모델 매트릭스",
+  "admin.models.missingConfig": "AI 등급 구성을 찾을 수 없습니다. 페이지를 새로 고침하세요.",
+  "admin.models.moveDown": "아래로 이동",
+  "admin.models.moveUp": "위로 이동",
+  "admin.models.networkDesc": "백엔드에 연결할 수 없습니다.",
+  "admin.models.networkTitle": "네트워크 오류",
+  "admin.models.premiumBadge": "프리미엄 AI 등급",
+  "admin.models.premiumCost": "— 비용 계수 5로 청구",
+  "admin.models.premiumReservedNote": "Business·Enterprise 플랜 전용 («openrouter_premium» 엔진).",
+  "admin.models.proBadge": "Pro 플랜",
+  "admin.models.remove": "제거",
+  "admin.models.save": "저장",
+  "admin.models.saveErrorTitle": "저장 오류",
+  "admin.models.saveNetworkDesc": "구성을 저장할 수 없습니다.",
+  "admin.models.savedDesc": "새 기본 모델은 다음 번역부터 재배포 없이 적용됩니다.",
+  "admin.models.savedTitle": "모델 저장됨",
+  "admin.models.saving": "저장 중...",
+  "admin.models.setDefault": "{model}을(를) 기본 모델로 설정",
+  "admin.models.sharedTierNote": "공유 등급: Business 플랜도 «openrouter» 엔진에는 에센셜 등급을 사용합니다 — 위 목록은 두 플랜 모두에 적용됩니다.",
+  "admin.models.subtitle": "플랜 → AI 등급 매트릭스. 번역 시 실제 사용되는 모델은 플랜 등급의 모델입니다: Pro 플랜은 프리미엄 모델을 절대 호출할 수 없습니다.",
+  "admin.models.title": "모델 & 구독",
+  "admin.stats.aiTiers": "AI 등급 (30일)",
+  "admin.stats.includingCredits": "크레딧 {amount} 포함",
+  "admin.stats.mrr": "추정 MRR",
+  "admin.stats.noTranslationsYet": "최근 30일간 번역 기록이 없습니다: 앞으로의 번역으로 등급별 분포가 채워집니다.",
+  "admin.stats.payments30": "30일간 결제 {count}건",
+  "admin.stats.refreshing": "새로 고치는 중...",
+  "admin.stats.revenue30": "매출 (30일)",
+  "admin.stats.revenueTotal": "총 수령 매출",
+  "admin.stats.tiersSub": "에센셜 / 프리미엄 · 클래식 {classic} · 기타 {other}",
+  "admin.stats.unavailable": "비즈니스 통계를 사용할 수 없습니다 ({error}).",
+  "admin.stats.waitlist": "대기 명단",
+  "admin.stats.waitlistSub": "대기 중인 등록자"
 }
--- a/frontend/src/lib/i18n/messages/ko/landing.json
+++ b/frontend/src/lib/i18n/messages/ko/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "전문가용",
  "landing.pricing.pro.f1": "월 200개 문서",
  "landing.pricing.pro.f2": "문서당 최대 200페이지",
-  "landing.pricing.pro.f3": "AI 번역",
+  "landing.pricing.pro.f3": "에센셜 AI: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google 포함",
  "landing.pricing.pro.f5": "맞춤 용어집 및 프롬프트",
  "landing.pricing.pro.f6": "우선 지원",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "대용량이 필요한 팀용",
  "landing.pricing.business.f1": "월 1,000개 문서",
  "landing.pricing.business.f2": "문서당 최대 500페이지",
-  "landing.pricing.business.f3": "프리미엄 AI (Claude)",
+  "landing.pricing.business.f3": "프리미엄 AI (Claude 5)",
  "landing.pricing.business.f4": "모든 제공자 + API",
  "landing.pricing.business.f5": "웹훅 및 자동화",
  "landing.pricing.business.f6": "팀 시트 5석",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "DOCX, XLSX, PPTX 또는 PDF 파일 지원",
  "landing.translate.aiAnalysis": "AI 분석 활성",
  "landing.translate.processing": "처리 중",
-  "landing.translate.preservingLayout": "레이아웃이 보존되고 있습니다"
+  "landing.translate.preservingLayout": "레이아웃이 보존되고 있습니다",
+  "landing.beforeAfter.seal": "동일한 레이아웃, 단어 하나까지 그대로",
+  "landing.beforeAfter.proof1": "SmartArt 다이어그램은 병합되지 않고 재구성됩니다",
+  "landing.beforeAfter.proof2": "차트 계열과 축 번역",
+  "landing.beforeAfter.proof3": "목차는 대상 언어로 재생성",
+  "landing.beforeAfter.targetTitle": "기술 사양 — 공기 처리",
+  "landing.beforeAfter.targetPara1": "환기 설비는 2분기 종료 전에 가동되어야 합니다.",
+  "landing.beforeAfter.targetPara2Before": "",
+  "landing.beforeAfter.targetTerm": "공기조화기",
+  "landing.beforeAfter.targetPara2After": "는 일정에 따라 F7 필터 등급을 준수해야 합니다.",
+  "landing.beforeAfter.targetItem1": "열부하: 정격 유량에서 42kW",
+  "landing.beforeAfter.targetItem2": "소음 수준: 3m에서 45dB(A) 미만",
+  "landing.formats.pill": "호환성",
+  "landing.hero.visualCaption": "동일한 레이아웃, 새로운 언어 — 그 외에는 아무것도 바뀌지 않습니다",
+  "landing.pricing.free.name": "무료",
+  "landing.pricing.free.desc": "앱을 경험하기에 적합",
+  "landing.pricing.free.cta": "이 플랜 선택",
+  "landing.pricing.enterprise.name": "엔터프라이즈",
+  "landing.pricing.enterprise.desc": "대규모 조직을 위한 맞춤 솔루션",
+  "landing.pricing.enterprise.cta": "문의하기",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/ko/pricing.json
+++ b/frontend/src/lib/i18n/messages/ko/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "90일 기록",
  "pricing.plans.business.feat1": "월 1,000개 문서",
  "pricing.plans.business.feat2": "문서당 최대 500페이지",
-  "pricing.plans.business.feat3": "에센셜 + 프리미엄 AI (Claude Haiku)",
+  "pricing.plans.business.feat3": "에센셜 + 프리미엄 AI (Claude 5)",
  "pricing.plans.business.feat4": "모든 번역 제공업체",
  "pricing.plans.business.feat5": "최대 50 MB 파일",
  "pricing.plans.business.feat6": "API 액세스 (월 10,000회 호출)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "1년 기록",
  "pricing.plans.business.feat10": "고급 분석",
  "pricing.plans.enterprise.feat1": "무제한 문서",
-  "pricing.plans.enterprise.feat2": "모든 AI 모델 (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "모든 AI 모델 (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "온프레미스 또는 전용 클라우드 배포",
  "pricing.plans.enterprise.feat4": "99.9% SLA 보장",
  "pricing.plans.enterprise.feat5": "24/7 전담 지원",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "초고속 멀티스레드 AI",
  "pricing.trust.availability.title": "24/7 이용 가능",
  "pricing.trust.availability.sub": "99.9% 가동 보장",
-  "pricing.aiModels.title": "AI 모델 — 2026년 3월",
+  "pricing.aiModels.title": "AI 모델 — 2026년 9월",
  "pricing.aiModels.essential.title": "에센셜 AI 번역",
  "pricing.aiModels.essential.plan": "Pro 플랜",
  "pricing.aiModels.essential.descPrefix": "기반:",
-  "pricing.aiModels.essential.descSuffix": "— 2026년 가장 비용 효율적인 AI 모델. 프론티어 모델과 동등한 품질을 극소수의 비용으로 제공.",
-  "pricing.aiModels.essential.modelName": "Essential AI 모델",
-  "pricing.aiModels.essential.context": "163K 토큰 컨텍스트",
-  "pricing.aiModels.essential.value": "우수한 가성비",
+  "pricing.aiModels.essential.descSuffix": "— 2026년 가장 비용 효율적인 AI 모델.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
+  "pricing.aiModels.essential.context": "최대 130만 토큰 컨텍스트 (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "최고의 가성비",
  "pricing.aiModels.premium.title": "프리미엄 AI 번역",
  "pricing.aiModels.premium.plan": "Business 플랜",
  "pricing.aiModels.premium.descPrefix": "기반:",
  "pricing.aiModels.premium.descSuffix": "Anthropic사 제품 — 법률, 의료 및 복잡한 기술 문서에서 높은 정확도.",
-  "pricing.aiModels.premium.context": "200K 토큰 컨텍스트",
+  "pricing.aiModels.premium.context": "100만 토큰 컨텍스트",
  "pricing.aiModels.premium.precision": "최고 정확도",
  "pricing.faq.title": "자주 묻는 질문",
  "pricing.faq.q1": "언제든지 플랜을 변경할 수 있나요?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "「에센셜 AI 번역」이란 무엇인가요?",
  "pricing.faq.a2": "당사의 AI 엔진입니다. 문서의 맥락을 이해하고, 레이아웃을 보존하며, 기술 용어를 기존 번역보다 훨씬 잘 처리합니다.",
  "pricing.faq.q3": "에센셜 AI와 프리미엄 AI의 차이점은 무엇인가요?",
-  "pricing.faq.a3": "Essential AI는 최적화된 모델을 사용합니다 (우수한 가성비). Premium AI는 Anthropic의 Claude 3.5 Haiku를 사용하여 법률, 의료 및 복잡한 기술 문서에서 더 정확합니다.",
+  "pricing.faq.a3": "Essential AI는 DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3를 사용합니다 (우수한 가성비). Premium AI는 Anthropic의 Claude 5를 사용하여 법률, 의료 및 복잡한 기술 문서에서 더 정확합니다.",
  "pricing.faq.q4": "번역 후 문서가 보관되나요?",
  "pricing.faq.a4": "번역된 파일은 플랜에 따라 이용 가능합니다 (Starter 30일, Pro 90일, Business 1년). 저장 시 및 전송 중 암호화됩니다.",
  "pricing.faq.q5": "월간 할당량을 초과하면 어떻게 되나요?",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "결제 생성 중 오류가 발생했습니다.",
  "pricing.dashboard": "대시보드",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "백만 토큰당 $0.09부터",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "대체 모델: DeepSeek V4 Pro, GLM-5.3",
+  "pricing.confirm.title": "구독 확인",
+  "pricing.confirm.subtitle": "안전한 결제 대행사로 이동합니다.",
+  "pricing.confirm.cancel": "취소",
+  "pricing.confirm.cta": "결제로 계속",
+  "pricing.confirm.year": "년",
+  "pricing.confirm.month": "개월",
+  "pricing.confirm.monthlyEquivalent": "연간 결제 — 월 {price}유로 상당.",
+  "pricing.confirm.secureNote": "프로필에서 언제든지 해지할 수 있습니다. 결제는 Stripe에서 처리되며 카드 정보는 당사 서버에 저장되지 않습니다.",
+  "pricing.enterprise.subject": "엔터프라이즈 문의",
+  "pricing.enterpriseBand.cta": "문의하기",
+  "pricing.enterpriseBand.text": "대용량, 전용 엔진, 온프레미스 옵션 — 상담해 보세요.",
+  "pricing.error.server": "서버 오류 {status}",
+  "pricing.freeBand.cta": "무료로 시작",
+  "pricing.freeBand.text": "일단 써보고 싶으신가요? 무료로 시작 — 월 5개 문서, 카드 불필요.",
+  "pricing.header.titleBase": "모든 필요를 위한",
+  "pricing.header.titleAccent": "하나의 플랜",
+  "pricing.toast.close": "닫기",
+  "pricing.aiModels.premium.price": "백만 토큰당 $2 / $10"
 }
--- a/frontend/src/lib/i18n/messages/ko/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/ko/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "번역에 최적화된 최고의 오픈소스 모델에 대한 통합 액세스.",
  "providerTheme.openrouter_premium.badge": "울트라",
  "providerTheme.openrouter_premium.subBadge": "최대 컨텍스트",
-  "providerTheme.openrouter_premium.desc": "최신 모델(GPT-4o, Claude Sonnet 4.6)의 지원을 받는 긴 문서 번역.",
+  "providerTheme.openrouter_premium.desc": "최신 모델(GPT-4o, Claude 5)의 지원을 받는 긴 문서 번역.",
  "providerTheme.zai.badge": "전문",
  "providerTheme.zai.subBadge": "금융 및 법률",
  "providerTheme.zai.desc": "까다로운 비즈니스 용어(법률, 금융)에 맞게 미세 조정된 모델.",
--- a/frontend/src/lib/i18n/messages/nl/admin.json
+++ b/frontend/src/lib/i18n/messages/nl/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Providers",
  "admin.nav.system": "Systeem",
  "admin.nav.logs": "Logboeken",
+  "admin.nav.stats": "Statistieken",
  "admin.users.title": "Gebruikersbeheer",
  "admin.users.subtitle": "Gebruikersaccounts bekijken en beheren",
  "admin.users.planUpdated": "Abonnement bijgewerkt",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "Wachten op gegevens...",
  "admin.system.purging": "Wissen...",
  "admin.system.clean": "Opschonen",
-  "admin.system.purge": "Wissen"
+  "admin.system.purge": "Wissen",
+  "admin.nav.models": "Modellen & abonnementen",
+  "admin.nav.marketing": "Marketing",
+  "admin.marketing.audience": "Doelgroep",
+  "admin.marketing.audienceAllUsers": "Alle accounts",
+  "admin.marketing.audienceInactive": "Accounts 30 dagen inactief",
+  "admin.marketing.audiencePlaceholder": "Kies een doelgroep",
+  "admin.marketing.audiencePlanBusiness": "Business-abonnement",
+  "admin.marketing.audiencePlanEnterprise": "Enterprise-abonnement",
+  "admin.marketing.audiencePlanFree": "Gratis-abonnement",
+  "admin.marketing.audiencePlanPro": "Pro-abonnement",
+  "admin.marketing.audiencePlanStarter": "Starter-abonnement",
+  "admin.marketing.audienceWaitlist": "Wachtlijst",
+  "admin.marketing.badgeReal": "echt",
+  "admin.marketing.badgeTest": "test",
+  "admin.marketing.confirmSend": "Deze e-mail versturen naar {count} ontvanger(s) („{audience}”)?",
+  "admin.marketing.footerNote": "Een voettekst met de uitschrijflink wordt automatisch toegevoegd bij verzending.",
+  "admin.marketing.hidePreview": "Voorbeeld verbergen",
+  "admin.marketing.history": "Verzendgeschiedenis",
+  "admin.marketing.historyFailedSuffix": ", {count} mislukt",
+  "admin.marketing.historySent": "{sent}/{total} verzonden",
+  "admin.marketing.html": "HTML-inhoud",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Vul het onderwerp en de HTML in.",
+  "admin.marketing.incompleteTitle": "Onvolledige inhoud",
+  "admin.marketing.loadErrorTitle": "Laadfout",
+  "admin.marketing.localCheckUnavailable": "Lokale controle niet beschikbaar in deze browser: de server controleert de voorafgaande testzending.",
+  "admin.marketing.networkDesc": "Kan de backend niet bereiken.",
+  "admin.marketing.newCampaign": "Nieuwe campagne",
+  "admin.marketing.noHistory": "Geen zending geregistreerd.",
+  "admin.marketing.preview": "Voorbeeld",
+  "admin.marketing.previewEmpty": "Voorbeeld: plak je HTML hiernaast.",
+  "admin.marketing.previewTitle": "E-mailvoorbeeld",
+  "admin.marketing.recipientsAvailable": "{count} ontvanger(s) beschikbaar na uitsluiting van uitschrijvingen.",
+  "admin.marketing.sendReal": "Versturen naar {count} ontvanger(s)",
+  "admin.marketing.sendRefusedTitle": "Verzending geweigerd",
+  "admin.marketing.sendTest": "Testzending",
+  "admin.marketing.sendingDesc": "{count} e-mail(s) in de verzendwachtrij (interval 0,2 s). Uitschrijvingen worden uitgesloten.",
+  "admin.marketing.sendingTitle": "Verzending bezig",
+  "admin.marketing.subject": "Onderwerp",
+  "admin.marketing.subjectPlaceholder": "Bijv.: Uw vertaling wacht — 20% gedurende 7 dagen",
+  "admin.marketing.subtitle": "E-mails sturen naar een doelgroep, met verplichte testzending, automatische uitschrijflink en volledige geschiedenis.",
+  "admin.marketing.testDoneDesc": "E-mail ontvangen op {email}",
+  "admin.marketing.testDoneTitle": "Testzending uitgevoerd",
+  "admin.marketing.testEmail": "Test-e-mail (optioneel)",
+  "admin.marketing.testEmailPlaceholder": "Anders: SMTP-afzenderadres",
+  "admin.marketing.testFailedTitle": "Testzending mislukt",
+  "admin.marketing.testRequiredHint": "Een echte zending vereist een voorafgaande testzending van exact deze inhoud (door de server gecontroleerd).",
+  "admin.marketing.testValidated": "Testzending gevalideerd voor deze inhoud.",
+  "admin.marketing.title": "Marketing — e-mailcampagnes",
+  "admin.marketing.unsubscribedNote": "{count} uitgeschreven contact(en) worden van elke zending uitgesloten.",
+  "admin.models.addModel": "Model toevoegen",
+  "admin.models.businessBadge": "Business-abonnement",
+  "admin.models.cancel": "Annuleren",
+  "admin.models.catalog": "OpenRouter-catalogus",
+  "admin.models.catalogErrorDesc": "OpenRouter-catalogus kon niet worden geladen.",
+  "admin.models.catalogErrorTitle": "Catalogus niet beschikbaar",
+  "admin.models.customBadge": "Aangepast niveau",
+  "admin.models.customNote": "Maatwerkmodellen, per geval met de klant afgesproken.",
+  "admin.models.defaultBadge": "standaard",
+  "admin.models.emptyTier": "Geen model: de officiële modellenreeks van het abonnement wordt gebruikt.",
+  "admin.models.enterpriseBadge": "Enterprise-abonnement",
+  "admin.models.essentialBadge": "Basis-AI-niveau",
+  "admin.models.essentialCost": "— gefactoreerd met kostenfactor 1",
+  "admin.models.fallbackFirst": "(fallback nr. 1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — controleer je admin-token.",
+  "admin.models.loadErrorTitle": "Laadfout",
+  "admin.models.matrixDesc": "Actieve modellen per niveau — de volgorde van de lijst is de fallback-prioriteit. De keuzerondje kiest het standaardmodel, direct actief zonder heruitrol.",
+  "admin.models.matrixTitle": "Modelmatrix per abonnement",
+  "admin.models.missingConfig": "AI-niveauconfiguratie niet gevonden. Herlaad de pagina.",
+  "admin.models.moveDown": "Omlaag",
+  "admin.models.moveUp": "Omhoog",
+  "admin.models.networkDesc": "Kan de backend niet bereiken.",
+  "admin.models.networkTitle": "Netwerkfout",
+  "admin.models.premiumBadge": "Premium-AI-niveau",
+  "admin.models.premiumCost": "— gefactoreerd met kostenfactor 5",
+  "admin.models.premiumReservedNote": "Voorbehouden aan Business- en Enterprise-abonnementen (engine «openrouter_premium»).",
+  "admin.models.proBadge": "Pro-abonnement",
+  "admin.models.remove": "Verwijderen",
+  "admin.models.save": "Opslaan",
+  "admin.models.saveErrorTitle": "Opslagfout",
+  "admin.models.saveNetworkDesc": "Configuratie kon niet worden opgeslagen.",
+  "admin.models.savedDesc": "Het nieuwe standaardmodel wordt vanaf de volgende vertaling gebruikt, zonder heruitrol.",
+  "admin.models.savedTitle": "Modellen opgeslagen",
+  "admin.models.saving": "Opslaan...",
+  "admin.models.setDefault": "{model} als standaardmodel instellen",
+  "admin.models.sharedTierNote": "Gedeeld niveau: het Business-abonnement gebruikt het Basis-niveau ook voor zijn «openrouter»-engine — de bovenstaande lijst geldt voor beide abonnementen.",
+  "admin.models.subtitle": "Matrix abonnementen → AI-niveaus. Het model dat per vertaling echt wordt gebruikt is dat van het niveau van het abonnement: een Pro-abonnement kan nooit een Premium-model aanroepen.",
+  "admin.models.title": "Modellen & abonnementen",
+  "admin.stats.aiTiers": "AI-niveaus (30 d)",
+  "admin.stats.includingCredits": "waarvan {amount} aan credits",
+  "admin.stats.mrr": "Geschatte MRR",
+  "admin.stats.noTranslationsYet": "Geen vertalingen van de laatste 30 dagen in de database: de niveauverdeling vult zich aan met de komende vertalingen.",
+  "admin.stats.payments30": "{count} betaling(en) in 30 dagen",
+  "admin.stats.refreshing": "Vernieuwen...",
+  "admin.stats.revenue30": "Omzet (30 dagen)",
+  "admin.stats.revenueTotal": "Totaal ontvangen omzet",
+  "admin.stats.tiersSub": "Basis / Premium · klassiek {classic} · overig {other}",
+  "admin.stats.unavailable": "Bedrijfsstatistieken niet beschikbaar ({error}).",
+  "admin.stats.waitlist": "Wachtlijst",
+  "admin.stats.waitlistSub": "inschrijvingen op de wachtlijst"
 }
--- a/frontend/src/lib/i18n/messages/nl/landing.json
+++ b/frontend/src/lib/i18n/messages/nl/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "Voor veeleisende professionals",
  "landing.pricing.pro.f1": "200 documenten / maand",
  "landing.pricing.pro.f2": "Tot 200 pagina's per doc",
-  "landing.pricing.pro.f3": "IA-aangedreven vertaling",
+  "landing.pricing.pro.f3": "Basis-AI: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google inbegrepen",
  "landing.pricing.pro.f5": "Eigen glossaria en prompts",
  "landing.pricing.pro.f6": "Prioriteitsondersteuning",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "Voor teams met grote behoeften",
  "landing.pricing.business.f1": "1 000 documenten / maand",
  "landing.pricing.business.f2": "Tot 500 pagina's per doc",
-  "landing.pricing.business.f3": "Premium AI (Claude)",
+  "landing.pricing.business.f3": "Premium-AI (Claude 5)",
  "landing.pricing.business.f4": "Alle providers + API-toegang",
  "landing.pricing.business.f5": "Webhooks en automatisering",
  "landing.pricing.business.f6": "5 teamplekken",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "DOCX, XLSX, PPTX of PDF bestanden ondersteund",
  "landing.translate.aiAnalysis": "Actieve AI-analyse",
  "landing.translate.processing": "Verwerking bezig",
-  "landing.translate.preservingLayout": "Uw opmaak wordt behouden"
+  "landing.translate.preservingLayout": "Uw opmaak wordt behouden",
+  "landing.beforeAfter.seal": "Zelfde lay-out, woord voor woord",
+  "landing.beforeAfter.proof1": "SmartArt-diagrammen worden herbouwd, niet afgevlakt",
+  "landing.beforeAfter.proof2": "Diagramreeksen en assen vertaald",
+  "landing.beforeAfter.proof3": "Inhoudsopgaven opnieuw gegenereerd in de doeltaal",
+  "landing.beforeAfter.targetTitle": "Technische specificatie — Luchtbehandeling",
+  "landing.beforeAfter.targetPara1": "Het ventilatiesysteem moet vóór het einde van het tweede kwartaal in bedrijf worden gesteld.",
+  "landing.beforeAfter.targetPara2Before": "De",
+  "landing.beforeAfter.targetTerm": "luchtbehandelingsgroep",
+  "landing.beforeAfter.targetPara2After": "moet volgens planning voldoen aan filterklasse F7.",
+  "landing.beforeAfter.targetItem1": "Thermische belasting: 42 kW bij nominale doorstroom",
+  "landing.beforeAfter.targetItem2": "Geluidsniveau onder 45 dB(A) op 3 m",
+  "landing.formats.pill": "COMPATIBILITEIT",
+  "landing.hero.visualCaption": "Zelfde lay-out, nieuwe taal — verder verandert niets",
+  "landing.pricing.free.name": "Gratis",
+  "landing.pricing.free.desc": "Ideaal om de app te ontdekken",
+  "landing.pricing.free.cta": "Dit plan kiezen",
+  "landing.pricing.enterprise.name": "Enterprise",
+  "landing.pricing.enterprise.desc": "Op maat gemaakte oplossingen voor grote organisaties",
+  "landing.pricing.enterprise.cta": "Neem contact op",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/nl/pricing.json
+++ b/frontend/src/lib/i18n/messages/nl/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "90 dagen geschiedenis",
  "pricing.plans.business.feat1": "1 000 documenten / maand",
  "pricing.plans.business.feat2": "Tot 500 pagina's per document",
-  "pricing.plans.business.feat3": "Basis- + Premium-AI (Claude Haiku)",
+  "pricing.plans.business.feat3": "Basis-AI + Premium-AI (Claude 5)",
  "pricing.plans.business.feat4": "Alle vertaalproviders",
  "pricing.plans.business.feat5": "Bestanden tot 50 MB",
  "pricing.plans.business.feat6": "API-toegang (10 000 aanroepen/maand)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "1 jaar geschiedenis",
  "pricing.plans.business.feat10": "Geavanceerde analyses",
  "pricing.plans.enterprise.feat1": "Onbeperkte documenten",
-  "pricing.plans.enterprise.feat2": "Alle AI-modellen (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "Alle AI-modellen (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "On-premise of dedicated cloud",
  "pricing.plans.enterprise.feat4": "99,9 % SLA gegarandeerd",
  "pricing.plans.enterprise.feat5": "24/7 dedicated support",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "Ultrasnelle multi-threaded AI",
  "pricing.trust.availability.title": "24/7 beschikbaar",
  "pricing.trust.availability.sub": "99,9 % gegarandeerde beschikbaarheid",
-  "pricing.aiModels.title": "Onze AI-modellen — maart 2026",
+  "pricing.aiModels.title": "Onze AI-modellen — september 2026",
  "pricing.aiModels.essential.title": "AI-basisvertaling",
  "pricing.aiModels.essential.plan": "Pro-abonnement",
  "pricing.aiModels.essential.descPrefix": "Gebaseerd op",
-  "pricing.aiModels.essential.descSuffix": "— het meest kostenefficiënte AI-model van 2026. Kwaliteit vergelijkbaar met frontier-modellen voor een fractie van de kosten.",
-  "pricing.aiModels.essential.modelName": "ons Essential AI-model",
-  "pricing.aiModels.essential.context": "163K tokens context",
-  "pricing.aiModels.essential.value": "Uitstekende prijs-kwaliteitverhouding",
+  "pricing.aiModels.essential.descSuffix": "— de AI-modellen met de beste prijs-kwaliteitverhouding van 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash en MiniMax M3",
+  "pricing.aiModels.essential.context": "Tot 1,3 mln context (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "De beste prijs-kwaliteitverhouding",
  "pricing.aiModels.premium.title": "AI-premiumvertaling",
  "pricing.aiModels.premium.plan": "Business-abonnement",
  "pricing.aiModels.premium.descPrefix": "Gebaseerd op",
  "pricing.aiModels.premium.descSuffix": "van Anthropic — nauwkeurig bij juridische, medische en complexe technische documenten.",
-  "pricing.aiModels.premium.context": "200K tokens context",
+  "pricing.aiModels.premium.context": "1 mln context",
  "pricing.aiModels.premium.precision": "Beste nauwkeurigheid",
  "pricing.faq.title": "Veelgestelde vragen",
  "pricing.faq.q1": "Kan ik op elk moment van abonnement wisselen?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "Wat is «AI-basisvertaling»?",
  "pricing.faq.a2": "Het is onze IA-engine. Hij begrijpt de context van uw documenten, behoudt de lay-out en behandelt technische termen veel beter dan klassieke vertaling.",
  "pricing.faq.q3": "Wat is het verschil tussen Basis- en Premium-AI?",
-  "pricing.faq.a3": "Essential IA gebruikt een geoptimaliseerd model (uitstekende prijs-kwaliteitverhouding). Premium IA gebruikt Claude 3.5 Haiku van Anthropic, nauwkeuriger bij juridische, medische en complexe technische documenten.",
+  "pricing.faq.a3": "Basis-AI draait op DeepSeek V4 Flash, GLM-5.3 Flash en MiniMax M3 (uitstekende prijs-kwaliteitverhouding). Premium-AI gebruikt Claude 5 van Anthropic, nauwkeuriger bij juridische, medische en complexe technische documenten.",
  "pricing.faq.q4": "Worden mijn documenten bewaard na vertaling?",
  "pricing.faq.a4": "Vertaalde bestanden zijn beschikbaar volgens uw abonnement (30 dagen Starter, 90 dagen Pro, 1 jaar Business). Ze zijn versleuteld at-rest en in transit.",
  "pricing.faq.q5": "Wat gebeurt er als ik mijn maandelijkse quotum overschrijd?",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "Fout bij aanmaken van de betaling.",
  "pricing.dashboard": "Dashboard",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "Vanaf $0,09 per miljoen tokens",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Alternatieven: DeepSeek V4 Pro en GLM-5.3",
+  "pricing.confirm.title": "Bevestig uw abonnement",
+  "pricing.confirm.subtitle": "U wordt doorgestuurd naar onze veilige betaalprovider.",
+  "pricing.confirm.cancel": "Annuleren",
+  "pricing.confirm.cta": "Verder naar betaling",
+  "pricing.confirm.year": "jaar",
+  "pricing.confirm.month": "maand",
+  "pricing.confirm.monthlyEquivalent": "Jaarlijkse facturering — dus {price} € / maand.",
+  "pricing.confirm.secureNote": "Op elk moment opzegbaar via uw profiel. Betaling wordt verwerkt door Stripe; uw kaartgegevens raken onze servers nooit.",
+  "pricing.enterprise.subject": "Enterprise-aanvraag",
+  "pricing.enterpriseBand.cta": "Neem contact op",
+  "pricing.enterpriseBand.text": "Volume, speciale engines, on-premise opties — laten we praten.",
+  "pricing.error.server": "Serverfout {status}",
+  "pricing.freeBand.cta": "Gratis starten",
+  "pricing.freeBand.text": "Eerst uitproberen? Start gratis — 5 documenten per maand, geen kaart nodig.",
+  "pricing.header.titleBase": "Een plan voor",
+  "pricing.header.titleAccent": "elke behoefte",
+  "pricing.toast.close": "Sluiten",
+  "pricing.aiModels.premium.price": "$2 / $10 per 1M tokens"
 }
--- a/frontend/src/lib/i18n/messages/nl/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/nl/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Uniforme toegang tot de beste open-source modellen, geoptimaliseerd voor vertaling.",
  "providerTheme.openrouter_premium.badge": "Ultra",
  "providerTheme.openrouter_premium.subBadge": "Maximale context",
-  "providerTheme.openrouter_premium.desc": "Ondersteund door state-of-the-art modellen (GPT-4o, Claude Sonnet 4.6) voor lange documenten.",
+  "providerTheme.openrouter_premium.desc": "Ondersteund door state-of-the-art modellen (GPT-4o, Claude 5) voor lange documenten.",
  "providerTheme.zai.badge": "Gespecialiseerd",
  "providerTheme.zai.subBadge": "Financiën & Recht",
  "providerTheme.zai.desc": "Model afgestemd op veeleisende bedrijfsterminologieën (juridisch, financieel).",
--- a/frontend/src/lib/i18n/messages/pt/admin.json
+++ b/frontend/src/lib/i18n/messages/pt/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Provedores",
  "admin.nav.system": "Sistema",
  "admin.nav.logs": "Registros",
+  "admin.nav.stats": "Estatísticas",
  "admin.users.title": "Gestão de Usuários",
  "admin.users.subtitle": "Visualizar e gerenciar contas de usuários",
  "admin.users.planUpdated": "Plano atualizado",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "Aguardando dados...",
  "admin.system.purging": "Limpando...",
  "admin.system.clean": "Limpar",
-  "admin.system.purge": "Limpar"
+  "admin.system.purge": "Limpar",
+  "admin.nav.models": "Modelos e assinaturas",
+  "admin.nav.marketing": "Marketing",
+  "admin.marketing.audience": "Público",
+  "admin.marketing.audienceAllUsers": "Todas as contas",
+  "admin.marketing.audienceInactive": "Contas inativas há 30 dias",
+  "admin.marketing.audiencePlaceholder": "Escolher um público",
+  "admin.marketing.audiencePlanBusiness": "Plano Business",
+  "admin.marketing.audiencePlanEnterprise": "Plano Enterprise",
+  "admin.marketing.audiencePlanFree": "Plano Gratuito",
+  "admin.marketing.audiencePlanPro": "Plano Pro",
+  "admin.marketing.audiencePlanStarter": "Plano Starter",
+  "admin.marketing.audienceWaitlist": "Lista de espera",
+  "admin.marketing.badgeReal": "real",
+  "admin.marketing.badgeTest": "teste",
+  "admin.marketing.confirmSend": "Enviar este email para {count} destinatário(s) («{audience}»)?",
+  "admin.marketing.footerNote": "Um rodapé com o link de descadastro é adicionado automaticamente ao envio.",
+  "admin.marketing.hidePreview": "Ocultar pré-visualização",
+  "admin.marketing.history": "Histórico de envios",
+  "admin.marketing.historyFailedSuffix": ", {count} falhas",
+  "admin.marketing.historySent": "{sent}/{total} enviados",
+  "admin.marketing.html": "Conteúdo HTML",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Preencha o assunto e o HTML.",
+  "admin.marketing.incompleteTitle": "Conteúdo incompleto",
+  "admin.marketing.loadErrorTitle": "Erro ao carregar",
+  "admin.marketing.localCheckUnavailable": "Verificação local indisponível neste navegador: o servidor verificará o envio de teste prévio.",
+  "admin.marketing.networkDesc": "Não foi possível contatar o backend.",
+  "admin.marketing.newCampaign": "Novo envio",
+  "admin.marketing.noHistory": "Nenhum envio registrado.",
+  "admin.marketing.preview": "Pré-visualização",
+  "admin.marketing.previewEmpty": "Pré-visualização: cole seu HTML ao lado.",
+  "admin.marketing.previewTitle": "Pré-visualização do email",
+  "admin.marketing.recipientsAvailable": "{count} destinatário(s) disponíveis após excluir os descadastrados.",
+  "admin.marketing.sendReal": "Enviar para {count} destinatário(s)",
+  "admin.marketing.sendRefusedTitle": "Envio recusado",
+  "admin.marketing.sendTest": "Envio de teste",
+  "admin.marketing.sendingDesc": "{count} email(s) na fila de envio (intervalo de 0,2 s). Os descadastrados ficam excluídos.",
+  "admin.marketing.sendingTitle": "Envio em andamento",
+  "admin.marketing.subject": "Assunto",
+  "admin.marketing.subjectPlaceholder": "Ex.: Sua tradução o espera — 20% por 7 dias",
+  "admin.marketing.subtitle": "Envio de emails para um público, com envio de teste obrigatório, link de descadastro automático e histórico completo.",
+  "admin.marketing.testDoneDesc": "Email recebido em {email}",
+  "admin.marketing.testDoneTitle": "Envio de teste realizado",
+  "admin.marketing.testEmail": "Email de teste (opcional)",
+  "admin.marketing.testEmailPlaceholder": "Caso contrário: endereço de remetente SMTP",
+  "admin.marketing.testFailedTitle": "Falha no envio de teste",
+  "admin.marketing.testRequiredHint": "O envio real exige um envio de teste prévio deste conteúdo exato (verificado pelo servidor).",
+  "admin.marketing.testValidated": "Envio de teste validado para este conteúdo.",
+  "admin.marketing.title": "Marketing — envios por email",
+  "admin.marketing.unsubscribedNote": "{count} contato(s) descadastrado(s) serão excluídos de todo envio.",
+  "admin.models.addModel": "Adicionar um modelo",
+  "admin.models.businessBadge": "Plano Business",
+  "admin.models.cancel": "Cancelar",
+  "admin.models.catalog": "Catálogo do OpenRouter",
+  "admin.models.catalogErrorDesc": "Não foi possível carregar o catálogo do OpenRouter.",
+  "admin.models.catalogErrorTitle": "Catálogo indisponível",
+  "admin.models.customBadge": "Nível personalizado",
+  "admin.models.customNote": "Modelos sob medida, definidos caso a caso com o cliente.",
+  "admin.models.defaultBadge": "padrão",
+  "admin.models.emptyTier": "Nenhum modelo: a gama oficial do plano será usada.",
+  "admin.models.enterpriseBadge": "Plano Enterprise",
+  "admin.models.essentialBadge": "Nível de IA Essencial",
+  "admin.models.essentialCost": "— faturado com fator de custo 1",
+  "admin.models.fallbackFirst": "(reserva n.º 1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — verifique seu token de administrador.",
+  "admin.models.loadErrorTitle": "Erro ao carregar",
+  "admin.models.matrixDesc": "Modelos ativos por nível — a ordem da lista é a prioridade de reserva. O botão de opção escolhe o modelo padrão, aplicado imediatamente sem nova implantação.",
+  "admin.models.matrixTitle": "Matriz de modelos por plano",
+  "admin.models.missingConfig": "Configuração dos níveis de IA não encontrada. Recarregue a página.",
+  "admin.models.moveDown": "Mover para baixo",
+  "admin.models.moveUp": "Mover para cima",
+  "admin.models.networkDesc": "Não foi possível contatar o backend.",
+  "admin.models.networkTitle": "Erro de rede",
+  "admin.models.premiumBadge": "Nível de IA Premium",
+  "admin.models.premiumCost": "— faturado com fator de custo 5",
+  "admin.models.premiumReservedNote": "Reservado aos planos Business e Enterprise (motor «openrouter_premium»).",
+  "admin.models.proBadge": "Plano Pro",
+  "admin.models.remove": "Remover",
+  "admin.models.save": "Salvar",
+  "admin.models.saveErrorTitle": "Erro ao salvar",
+  "admin.models.saveNetworkDesc": "Não foi possível salvar a configuração.",
+  "admin.models.savedDesc": "O novo modelo padrão é usado já na próxima tradução, sem nova implantação.",
+  "admin.models.savedTitle": "Modelos salvos",
+  "admin.models.saving": "Salvando...",
+  "admin.models.setDefault": "Definir {model} como modelo padrão",
+  "admin.models.sharedTierNote": "Nível compartilhado: o plano Business também usa o Essencial para o motor «openrouter» — a lista acima vale para os dois planos.",
+  "admin.models.subtitle": "Matriz planos → níveis de IA. O modelo realmente usado em cada tradução é o do nível do plano: um plano Pro nunca pode disparar um modelo Premium.",
+  "admin.models.title": "Modelos e assinaturas",
+  "admin.stats.aiTiers": "Níveis de IA (30 d)",
+  "admin.stats.includingCredits": "incluindo {amount} em créditos",
+  "admin.stats.mrr": "MRR estimado",
+  "admin.stats.noTranslationsYet": "Nenhuma tradução registrada nos últimos 30 dias: a distribuição por níveis se preencherá com as próximas traduções.",
+  "admin.stats.payments30": "{count} pagamento(s) em 30 dias",
+  "admin.stats.refreshing": "Atualizando...",
+  "admin.stats.revenue30": "Receitas (30 dias)",
+  "admin.stats.revenueTotal": "Receitas recebidas (total)",
+  "admin.stats.tiersSub": "Essencial / Premium · clássico {classic} · outro {other}",
+  "admin.stats.unavailable": "Estatísticas de negócio indisponíveis ({error}).",
+  "admin.stats.waitlist": "Lista de espera",
+  "admin.stats.waitlistSub": "inscritos na espera"
 }
--- a/frontend/src/lib/i18n/messages/pt/landing.json
+++ b/frontend/src/lib/i18n/messages/pt/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "Para profissionais exigentes",
  "landing.pricing.pro.f1": "200 documentos / mês",
  "landing.pricing.pro.f2": "Até 200 páginas por doc",
-  "landing.pricing.pro.f3": "Tradução com IA",
+  "landing.pricing.pro.f3": "IA Essencial: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google incluídos",
  "landing.pricing.pro.f5": "Glossários e prompts personalizados",
  "landing.pricing.pro.f6": "Suporte prioritário",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "Para equipes com alto volume",
  "landing.pricing.business.f1": "1 000 documentos / mês",
  "landing.pricing.business.f2": "Até 500 páginas por doc",
-  "landing.pricing.business.f3": "IA Premium (Claude)",
+  "landing.pricing.business.f3": "IA Premium (Claude 5)",
  "landing.pricing.business.f4": "Todos os provedores + API",
  "landing.pricing.business.f5": "Webhooks e automação",
  "landing.pricing.business.f6": "5 assentos de equipe",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "Ficheiros DOCX, XLSX, PPTX ou PDF suportados",
  "landing.translate.aiAnalysis": "Análise IA Ativa",
  "landing.translate.processing": "Em processamento",
-  "landing.translate.preservingLayout": "O seu layout está a ser preservado"
+  "landing.translate.preservingLayout": "O seu layout está a ser preservado",
+  "landing.beforeAfter.seal": "O mesmo layout, palavra por palavra",
+  "landing.beforeAfter.proof1": "Diagramas SmartArt reconstruídos, não achatados",
+  "landing.beforeAfter.proof2": "Séries e eixos dos gráficos traduzidos",
+  "landing.beforeAfter.proof3": "Sumários regenerados no idioma de destino",
+  "landing.beforeAfter.targetTitle": "Especificação técnica — Tratamento de ar",
+  "landing.beforeAfter.targetPara1": "A instalação de ventilação deve ser colocada em serviço antes do fim do segundo trimestre.",
+  "landing.beforeAfter.targetPara2Before": "A",
+  "landing.beforeAfter.targetTerm": "unidade de tratamento de ar",
+  "landing.beforeAfter.targetPara2After": "deve atender à classe de filtro F7 conforme o cronograma.",
+  "landing.beforeAfter.targetItem1": "Carga térmica: 42 kW em vazão nominal",
+  "landing.beforeAfter.targetItem2": "Nível sonoro abaixo de 45 dB(A) a 3 m",
+  "landing.formats.pill": "COMPATIBILIDADE",
+  "landing.hero.visualCaption": "O mesmo layout, novo idioma — nada mais muda",
+  "landing.pricing.free.name": "Grátis",
+  "landing.pricing.free.desc": "Ideal para descobrir a aplicação",
+  "landing.pricing.free.cta": "Escolher este plano",
+  "landing.pricing.enterprise.name": "Empresas",
+  "landing.pricing.enterprise.desc": "Soluções à medida para grandes organizações",
+  "landing.pricing.enterprise.cta": "Fale conosco",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/pt/pricing.json
+++ b/frontend/src/lib/i18n/messages/pt/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "Histórico de 90 dias",
  "pricing.plans.business.feat1": "1 000 documentos / mês",
  "pricing.plans.business.feat2": "Até 500 páginas por documento",
-  "pricing.plans.business.feat3": "IA Essencial + Premium (Claude Haiku)",
+  "pricing.plans.business.feat3": "IA Essencial + IA Premium (Claude 5)",
  "pricing.plans.business.feat4": "Todos os provedores de tradução",
  "pricing.plans.business.feat5": "Arquivos de até 50 MB",
  "pricing.plans.business.feat6": "Acesso à API (10 000 chamadas/mês)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "Histórico de 1 ano",
  "pricing.plans.business.feat10": "Análises avançadas",
  "pricing.plans.enterprise.feat1": "Documentos ilimitados",
-  "pricing.plans.enterprise.feat2": "Todos os modelos de IA (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "Todos os modelos de IA (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "Implantação local ou em nuvem dedicada",
  "pricing.plans.enterprise.feat4": "SLA 99,9 % garantido",
  "pricing.plans.enterprise.feat5": "Suporte dedicado 24/7",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "IA multi-thread ultrarrápida",
  "pricing.trust.availability.title": "Disponível 24/7",
  "pricing.trust.availability.sub": "99,9 % de uptime garantido",
-  "pricing.aiModels.title": "Nossos modelos de IA — Março 2026",
+  "pricing.aiModels.title": "Nossos modelos de IA — Setembro de 2026",
  "pricing.aiModels.essential.title": "Tradução IA Essencial",
  "pricing.aiModels.essential.plan": "Plano Pro",
  "pricing.aiModels.essential.descPrefix": "Baseado em",
-  "pricing.aiModels.essential.descSuffix": "— o modelo de IA mais econômico de 2026. Qualidade comparável a modelos frontier a uma fração do custo.",
-  "pricing.aiModels.essential.modelName": "nosso modelo IA Essencial",
-  "pricing.aiModels.essential.context": "163K tokens de contexto",
-  "pricing.aiModels.essential.value": "Excelente custo-benefício",
+  "pricing.aiModels.essential.descSuffix": "— os modelos de IA com o melhor custo-benefício de 2026.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash e MiniMax M3",
+  "pricing.aiModels.essential.context": "Até 1,3 M de contexto (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "O melhor custo-benefício",
  "pricing.aiModels.premium.title": "Tradução IA Premium",
  "pricing.aiModels.premium.plan": "Plano Business",
  "pricing.aiModels.premium.descPrefix": "Baseado em",
  "pricing.aiModels.premium.descSuffix": "da Anthropic — preciso em documentos jurídicos, médicos e técnicos complexos.",
-  "pricing.aiModels.premium.context": "200K tokens de contexto",
+  "pricing.aiModels.premium.context": "1M de contexto",
  "pricing.aiModels.premium.precision": "Maior precisão",
  "pricing.faq.title": "Perguntas frequentes",
  "pricing.faq.q1": "Posso trocar de plano a qualquer momento?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "O que é a «Tradução IA Essencial»?",
  "pricing.faq.a2": "É nosso motor de IA. Ele compreende o contexto dos seus documentos, preserva a formatação e lida com termos técnicos muito melhor do que a tradução clássica.",
  "pricing.faq.q3": "Qual a diferença entre IA Essencial e IA Premium?",
-  "pricing.faq.a3": "A IA Essencial usa um modelo otimizado (excelente custo-benefício). A IA Premium usa Claude 3.5 Haiku da Anthropic, mais precisa em documentos jurídicos, médicos e técnicos complexos.",
+  "pricing.faq.a3": "A IA Essencial usa DeepSeek V4 Flash, GLM-5.3 Flash e MiniMax M3 (excelente custo-benefício). A IA Premium usa Claude 5 da Anthropic, mais precisa em documentos jurídicos, médicos e técnicos complexos.",
  "pricing.faq.q4": "Meus documentos são mantidos após a tradução?",
  "pricing.faq.a4": "Os arquivos traduzidos ficam disponíveis de acordo com seu plano (30 dias Starter, 90 dias Pro, 1 ano Business). São criptografados em repouso e em trânsito.",
  "pricing.faq.q5": "O que acontece se eu exceder minha cota mensal?",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "Erro ao criar o pagamento.",
  "pricing.dashboard": "Painel",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "A partir de 0,09 US$ por milhão de tokens",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Alternativas: DeepSeek V4 Pro e GLM-5.3",
+  "pricing.confirm.title": "Confirme a sua assinatura",
+  "pricing.confirm.subtitle": "Você será redirecionado ao nosso provedor de pagamento seguro.",
+  "pricing.confirm.cancel": "Cancelar",
+  "pricing.confirm.cta": "Continuar para o pagamento",
+  "pricing.confirm.year": "ano",
+  "pricing.confirm.month": "mês",
+  "pricing.confirm.monthlyEquivalent": "Faturação anual — ou seja, {price} € / mês.",
+  "pricing.confirm.secureNote": "Cancelável a qualquer momento no seu perfil. O pagamento é processado pelo Stripe; os dados do seu cartão nunca passam pelos nossos servidores.",
+  "pricing.enterprise.subject": "Pedido Enterprise",
+  "pricing.enterpriseBand.cta": "Fale conosco",
+  "pricing.enterpriseBand.text": "Volume, motores dedicados, opções on-premise — vamos conversar.",
+  "pricing.error.server": "Erro do servidor {status}",
+  "pricing.freeBand.cta": "Comece grátis",
+  "pricing.freeBand.text": "Só quer experimentar? Comece grátis — 5 documentos por mês, sem cartão.",
+  "pricing.header.titleBase": "Um plano para",
+  "pricing.header.titleAccent": "cada necessidade",
+  "pricing.toast.close": "Fechar",
+  "pricing.aiModels.premium.price": "2 US$ / 10 US$ por 1M de tokens"
 }
--- a/frontend/src/lib/i18n/messages/pt/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/pt/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Acesso unificado aos melhores modelos open-source otimizados para tradução.",
  "providerTheme.openrouter_premium.badge": "Ultra",
  "providerTheme.openrouter_premium.subBadge": "Contexto máximo",
-  "providerTheme.openrouter_premium.desc": "Assistido pelos modelos mais avançados (GPT-4o, Claude Sonnet 4.6) para documentos longos.",
+  "providerTheme.openrouter_premium.desc": "Assistido pelos modelos mais avançados (GPT-4o, Claude 5) para documentos longos.",
  "providerTheme.zai.badge": "Especializada",
  "providerTheme.zai.subBadge": "Finanças e Direito",
  "providerTheme.zai.desc": "Modelo afinado para terminologias empresariais exigentes (jurídico, financeiro).",
--- a/frontend/src/lib/i18n/messages/ru/admin.json
+++ b/frontend/src/lib/i18n/messages/ru/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "Провайдеры",
  "admin.nav.system": "Система",
  "admin.nav.logs": "Журналы",
+  "admin.nav.stats": "Статистика",
  "admin.users.title": "Управление пользователями",
  "admin.users.subtitle": "Просмотр и управление учётными записями",
  "admin.users.planUpdated": "План обновлён",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "Ожидание данных...",
  "admin.system.purging": "Очистка...",
  "admin.system.clean": "Очистить",
-  "admin.system.purge": "Очистить"
+  "admin.system.purge": "Очистить",
+  "admin.nav.models": "Модели и подписки",
+  "admin.nav.marketing": "Маркетинг",
+  "admin.marketing.audience": "Аудитория",
+  "admin.marketing.audienceAllUsers": "Все аккаунты",
+  "admin.marketing.audienceInactive": "Аккаунты без активности 30 дней",
+  "admin.marketing.audiencePlaceholder": "Выберите аудиторию",
+  "admin.marketing.audiencePlanBusiness": "Тариф Business",
+  "admin.marketing.audiencePlanEnterprise": "Тариф Enterprise",
+  "admin.marketing.audiencePlanFree": "Бесплатный тариф",
+  "admin.marketing.audiencePlanPro": "Тариф Pro",
+  "admin.marketing.audiencePlanStarter": "Тариф Starter",
+  "admin.marketing.audienceWaitlist": "Лист ожидания",
+  "admin.marketing.badgeReal": "реальная",
+  "admin.marketing.badgeTest": "тест",
+  "admin.marketing.confirmSend": "Отправить это письмо {count} получателю(ям) («{audience}»)?",
+  "admin.marketing.footerNote": "При отправке автоматически добавляется подвал со ссылкой отписки.",
+  "admin.marketing.hidePreview": "Скрыть предпросмотр",
+  "admin.marketing.history": "История рассылок",
+  "admin.marketing.historyFailedSuffix": ", {count} сбоев",
+  "admin.marketing.historySent": "{sent}/{total} отправлено",
+  "admin.marketing.html": "HTML-содержимое",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "Заполните тему и HTML.",
+  "admin.marketing.incompleteTitle": "Содержимое неполное",
+  "admin.marketing.loadErrorTitle": "Ошибка загрузки",
+  "admin.marketing.localCheckUnavailable": "Локальная проверка в этом браузере недоступна: сервер проверит предварительную тестовую отправку.",
+  "admin.marketing.networkDesc": "Не удаётся связаться с сервером.",
+  "admin.marketing.newCampaign": "Новая рассылка",
+  "admin.marketing.noHistory": "Рассылок не зарегистрировано.",
+  "admin.marketing.preview": "Предпросмотр",
+  "admin.marketing.previewEmpty": "Предпросмотр: вставьте HTML рядом с этой панелью.",
+  "admin.marketing.previewTitle": "Предпросмотр письма",
+  "admin.marketing.recipientsAvailable": "{count} получателей доступно после исключения отписавшихся.",
+  "admin.marketing.sendReal": "Отправить {count} получателю(ям)",
+  "admin.marketing.sendRefusedTitle": "Рассылка отклонена",
+  "admin.marketing.sendTest": "Тестовая отправка",
+  "admin.marketing.sendingDesc": "{count} письмо(ем) в очереди отправки (интервал 0,2 с). Отписавшиеся исключаются.",
+  "admin.marketing.sendingTitle": "Идёт рассылка",
+  "admin.marketing.subject": "Тема",
+  "admin.marketing.subjectPlaceholder": "Напр.: Ваш перевод готов — скидка 20% на 7 дней",
+  "admin.marketing.subtitle": "Рассылка писем по аудитории с обязательной тестовой отправкой, автоматической ссылкой отписки и полной историей.",
+  "admin.marketing.testDoneDesc": "Письмо получено на {email}",
+  "admin.marketing.testDoneTitle": "Тестовая отправка выполнена",
+  "admin.marketing.testEmail": "Тестовый email (необязательно)",
+  "admin.marketing.testEmailPlaceholder": "Иначе: SMTP-адрес отправителя",
+  "admin.marketing.testFailedTitle": "Тестовая отправка не удалась",
+  "admin.marketing.testRequiredHint": "Реальной рассылке предшествует тестовая отправка точно такого же содержимого (проверяется сервером).",
+  "admin.marketing.testValidated": "Тестовая отправка этого содержимого подтверждена.",
+  "admin.marketing.title": "Маркетинг — email-рассылки",
+  "admin.marketing.unsubscribedNote": "{count} отписавшихся будут исключены из любой рассылки.",
+  "admin.models.addModel": "Добавить модель",
+  "admin.models.businessBadge": "Тариф Business",
+  "admin.models.cancel": "Отмена",
+  "admin.models.catalog": "Каталог OpenRouter",
+  "admin.models.catalogErrorDesc": "Не удалось загрузить каталог OpenRouter.",
+  "admin.models.catalogErrorTitle": "Каталог недоступен",
+  "admin.models.customBadge": "Индивидуальный уровень",
+  "admin.models.customNote": "Специальные модели, согласовываются с клиентом индивидуально.",
+  "admin.models.defaultBadge": "по умолчанию",
+  "admin.models.emptyTier": "Нет моделей: будет использован официальный набор тарифа.",
+  "admin.models.enterpriseBadge": "Тариф Enterprise",
+  "admin.models.essentialBadge": "Уровень ИИ «Базовый»",
+  "admin.models.essentialCost": "— с коэффициентом стоимости 1",
+  "admin.models.fallbackFirst": "(запасная № 1)",
+  "admin.models.loadErrorDesc": "HTTP {status} — проверьте токен администратора.",
+  "admin.models.loadErrorTitle": "Ошибка загрузки",
+  "admin.models.matrixDesc": "Активные модели по уровням — порядок в списке задаёт приоритет запасного варианта. Радиокнопка выбирает модель по умолчанию, применяемую сразу, без повторного развёртывания.",
+  "admin.models.matrixTitle": "Матрица моделей по тарифам",
+  "admin.models.missingConfig": "Конфигурация уровней ИИ не найдена. Перезагрузите страницу.",
+  "admin.models.moveDown": "Вниз",
+  "admin.models.moveUp": "Вверх",
+  "admin.models.networkDesc": "Не удаётся связаться с сервером.",
+  "admin.models.networkTitle": "Сетевая ошибка",
+  "admin.models.premiumBadge": "Уровень ИИ «Премиум»",
+  "admin.models.premiumCost": "— с коэффициентом стоимости 5",
+  "admin.models.premiumReservedNote": "Только для тарифов Business и Enterprise (движок «openrouter_premium»).",
+  "admin.models.proBadge": "Тариф Pro",
+  "admin.models.remove": "Убрать",
+  "admin.models.save": "Сохранить",
+  "admin.models.saveErrorTitle": "Ошибка сохранения",
+  "admin.models.saveNetworkDesc": "Не удалось сохранить конфигурацию.",
+  "admin.models.savedDesc": "Новая модель по умолчанию применяется уже со следующего перевода, без повторного развёртывания.",
+  "admin.models.savedTitle": "Модели сохранены",
+  "admin.models.saving": "Сохранение...",
+  "admin.models.setDefault": "Назначить {model} моделью по умолчанию",
+  "admin.models.sharedTierNote": "Общий уровень: тариф Business также использует «Базовый» для движка «openrouter» — приведённый выше список относится к обоим тарифам.",
+  "admin.models.subtitle": "Матрица тарифы → уровни ИИ. Модель, реально используемая при переводе, — это модель уровня тарифа: тариф Pro никогда не задействует модель Premium.",
+  "admin.models.title": "Модели и подписки",
+  "admin.stats.aiTiers": "Уровни ИИ (30 дн.)",
+  "admin.stats.includingCredits": "в том числе {amount} кредитами",
+  "admin.stats.mrr": "Оценочный MRR",
+  "admin.stats.noTranslationsYet": "За последние 30 дней переводов в базе нет: распределение по уровням заполнится следующими переводами.",
+  "admin.stats.payments30": "{count} платёж(ей) за 30 дней",
+  "admin.stats.refreshing": "Обновление...",
+  "admin.stats.revenue30": "Выручка (30 дней)",
+  "admin.stats.revenueTotal": "Всего получено выручки",
+  "admin.stats.tiersSub": "Базовый / Премиум · классика {classic} · прочее {other}",
+  "admin.stats.unavailable": "Бизнес-статистика недоступна ({error}).",
+  "admin.stats.waitlist": "Лист ожидания",
+  "admin.stats.waitlistSub": "записей в листе ожидания"
 }
--- a/frontend/src/lib/i18n/messages/ru/landing.json
+++ b/frontend/src/lib/i18n/messages/ru/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "Для требовательных профессионалов",
  "landing.pricing.pro.f1": "200 документов / мес.",
  "landing.pricing.pro.f2": "До 200 страниц на документ",
-  "landing.pricing.pro.f3": "Перевод на базе ИИ",
+  "landing.pricing.pro.f3": "Базовый ИИ: DeepSeek V4 Flash, GLM-5.3 Flash, MiniMax M3",
  "landing.pricing.pro.f4": "Google включены",
  "landing.pricing.pro.f5": "Пользовательские глоссарии и промпты",
  "landing.pricing.pro.f6": "Приоритетная поддержка",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "Для команд с большими объёмами",
  "landing.pricing.business.f1": "1 000 документов / мес.",
  "landing.pricing.business.f2": "До 500 страниц на документ",
-  "landing.pricing.business.f3": "Премиум ИИ (Claude)",
+  "landing.pricing.business.f3": "Премиум ИИ (Claude 5)",
  "landing.pricing.business.f4": "Все провайдеры + API",
  "landing.pricing.business.f5": "Вебхуки и автоматизация",
  "landing.pricing.business.f6": "5 рабочих мест",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "Поддерживаются файлы DOCX, XLSX, PPTX или PDF",
  "landing.translate.aiAnalysis": "Активный ИИ-анализ",
  "landing.translate.processing": "Обработка",
-  "landing.translate.preservingLayout": "Ваше форматирование сохраняется"
+  "landing.translate.preservingLayout": "Ваше форматирование сохраняется",
+  "landing.beforeAfter.seal": "Тот же макет, слово в слово",
+  "landing.beforeAfter.proof1": "Диаграммы SmartArt воссоздаются, а не сплющиваются",
+  "landing.beforeAfter.proof2": "Ряды и оси диаграмм переведены",
+  "landing.beforeAfter.proof3": "Оглавления заново формируются на целевом языке",
+  "landing.beforeAfter.targetTitle": "Техническая спецификация — Обработка воздуха",
+  "landing.beforeAfter.targetPara1": "Система вентиляции должна быть введена в эксплуатацию до конца второго квартала.",
+  "landing.beforeAfter.targetPara2Before": "",
+  "landing.beforeAfter.targetTerm": "Приточно-вытяжная установка",
+  "landing.beforeAfter.targetPara2After": "должна соответствовать классу фильтрации F7 согласно графику.",
+  "landing.beforeAfter.targetItem1": "Тепловая нагрузка: 42 кВт при номинальном расходе",
+  "landing.beforeAfter.targetItem2": "Уровень шума ниже 45 dB(A) на расстоянии 3 м",
+  "landing.formats.pill": "СОВМЕСТИМОСТЬ",
+  "landing.hero.visualCaption": "Тот же макет, новый язык — больше ничего не меняется",
+  "landing.pricing.free.name": "Бесплатно",
+  "landing.pricing.free.desc": "Идеально, чтобы попробовать приложение",
+  "landing.pricing.free.cta": "Выбрать этот план",
+  "landing.pricing.enterprise.name": "Enterprise",
+  "landing.pricing.enterprise.desc": "Индивидуальные решения для крупных организаций",
+  "landing.pricing.enterprise.cta": "Связаться с нами",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/ru/pricing.json
+++ b/frontend/src/lib/i18n/messages/ru/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "История за 90 дней",
  "pricing.plans.business.feat1": "1 000 документов / месяц",
  "pricing.plans.business.feat2": "До 500 страниц на документ",
-  "pricing.plans.business.feat3": "Базовый + Премиум ИИ (Claude Haiku)",
+  "pricing.plans.business.feat3": "Базовый + Премиум ИИ (Claude 5)",
  "pricing.plans.business.feat4": "Все провайдеры перевода",
  "pricing.plans.business.feat5": "Файлы до 50 МБ",
  "pricing.plans.business.feat6": "Доступ к API (10 000 вызовов/мес.)",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "История за 1 год",
  "pricing.plans.business.feat10": "Расширенная аналитика",
  "pricing.plans.enterprise.feat1": "Безлимитные документы",
-  "pricing.plans.enterprise.feat2": "Все ИИ-модели (GPT-5, Claude Opus 4.6…)",
+  "pricing.plans.enterprise.feat2": "Все ИИ-модели (Claude 5, DeepSeek V4 Pro, GLM-5.3…)",
  "pricing.plans.enterprise.feat3": "Локальное развёртывание или выделенное облако",
  "pricing.plans.enterprise.feat4": "SLA 99,9 % гарантировано",
  "pricing.plans.enterprise.feat5": "Выделенная поддержка 24/7",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "Сверхбыстрая многопоточная ИИ",
  "pricing.trust.availability.title": "Доступно 24/7",
  "pricing.trust.availability.sub": "99,9 % гарантированный аптайм",
-  "pricing.aiModels.title": "Наши ИИ-модели — март 2026",
+  "pricing.aiModels.title": "Наши ИИ-модели — сентябрь 2026",
  "pricing.aiModels.essential.title": "Базовый ИИ-перевод",
  "pricing.aiModels.essential.plan": "Тариф Pro",
  "pricing.aiModels.essential.descPrefix": "На базе",
-  "pricing.aiModels.essential.descSuffix": "— самой экономичной ИИ-модели 2026 года. Качество на уровне frontier-моделей при стоимости в десятки раз ниже.",
-  "pricing.aiModels.essential.modelName": "наша базовая ИИ-модель",
-  "pricing.aiModels.essential.context": "163K токенов контекста",
-  "pricing.aiModels.essential.value": "Отличное соотношение цены и качества",
+  "pricing.aiModels.essential.descSuffix": "— самые экономичные ИИ-модели 2026 года.",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash, GLM-5.3 Flash и MiniMax M3",
+  "pricing.aiModels.essential.context": "Контекст до 1,3 млн токенов (GLM-5.3 Flash)",
+  "pricing.aiModels.essential.value": "Лучшее соотношение цены и качества",
  "pricing.aiModels.premium.title": "Премиум ИИ-перевод",
  "pricing.aiModels.premium.plan": "Тариф Business",
  "pricing.aiModels.premium.descPrefix": "На базе",
  "pricing.aiModels.premium.descSuffix": "от Anthropic — высокая точность на юридических, медицинских и сложных технических документах.",
-  "pricing.aiModels.premium.context": "200K токенов контекста",
+  "pricing.aiModels.premium.context": "1 млн токенов контекста",
  "pricing.aiModels.premium.precision": "Наивысшая точность",
  "pricing.faq.title": "Часто задаваемые вопросы",
  "pricing.faq.q1": "Могу ли я сменить тариф в любое время?",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "Что такое «Базовый ИИ-перевод»?",
  "pricing.faq.a2": "Это наш ИИ-движок. Он понимает контекст ваших документов, сохраняет вёрстку и обрабатывает технические термины намного лучше классического перевода.",
  "pricing.faq.q3": "В чём разница между базовым и премиум ИИ-переводом?",
-  "pricing.faq.a3": "Базовый ИИ использует оптимизированную модель (отличное соотношение цены и качества). Премиум ИИ использует Claude 3.5 Haiku от Anthropic, более точный на юридических, медицинских и сложных технических документах.",
+  "pricing.faq.a3": "Базовый ИИ работает на моделях DeepSeek V4 Flash, GLM-5.3 Flash и MiniMax M3 (отличное соотношение цены и качества). Премиум ИИ использует Claude 5 от Anthropic, более точный на юридических, медицинских и сложных технических документах.",
  "pricing.faq.q4": "Сохраняются ли мои документы после перевода?",
  "pricing.faq.a4": "Переведённые файлы доступны в зависимости от тарифа (30 дней Starter, 90 дней Pro, 1 год Business). Они зашифрованы при хранении и передаче.",
  "pricing.faq.q5": "Что произойдёт при превышении месячной квоты?",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "Ошибка при создании платежа.",
  "pricing.dashboard": "Панель",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "От $0,09 за 1 млн токенов",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "Альтернативы: DeepSeek V4 Pro и GLM-5.3",
+  "pricing.confirm.title": "Подтвердите подписку",
+  "pricing.confirm.subtitle": "Вы будете перенаправлены к нашему безопасному платёжному провайдеру.",
+  "pricing.confirm.cancel": "Отмена",
+  "pricing.confirm.cta": "Перейти к оплате",
+  "pricing.confirm.year": "год",
+  "pricing.confirm.month": "месяц",
+  "pricing.confirm.monthlyEquivalent": "Годовая оплата — то есть {price} € / месяц.",
+  "pricing.confirm.secureNote": "Отмена в любой момент из профиля. Платёж обрабатывает Stripe; данные вашей карты никогда не попадают на наши серверы.",
+  "pricing.enterprise.subject": "Запрос по тарифу Enterprise",
+  "pricing.enterpriseBand.cta": "Связаться с нами",
+  "pricing.enterpriseBand.text": "Объёмы, выделенные движки, локальное размещение — давайте обсудим.",
+  "pricing.error.server": "Ошибка сервера {status}",
+  "pricing.freeBand.cta": "Начать бесплатно",
+  "pricing.freeBand.text": "Хотите просто попробовать? Начните бесплатно — 5 документов в месяц, без карты.",
+  "pricing.header.titleBase": "Тариф для",
+  "pricing.header.titleAccent": "каждой задачи",
+  "pricing.toast.close": "Закрыть",
+  "pricing.aiModels.premium.price": "$2 / $10 за 1 млн токенов"
 }
--- a/frontend/src/lib/i18n/messages/ru/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/ru/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "Единый доступ к лучшим open-source моделям, оптимизированным для перевода.",
  "providerTheme.openrouter_premium.badge": "Ультра",
  "providerTheme.openrouter_premium.subBadge": "Максимальный контекст",
-  "providerTheme.openrouter_premium.desc": "С помощью современных моделей (GPT-4o, Claude Sonnet 4.6) для длинных документов.",
+  "providerTheme.openrouter_premium.desc": "С помощью современных моделей (GPT-4o, Claude 5) для длинных документов.",
  "providerTheme.zai.badge": "Специализированный",
  "providerTheme.zai.subBadge": "Финансы и право",
  "providerTheme.zai.desc": "Модель точно настроена для требовательных бизнес-терминологий (юриспруденция, финансы).",
--- a/frontend/src/lib/i18n/messages/zh/admin.json
+++ b/frontend/src/lib/i18n/messages/zh/admin.json
@@ -20,6 +20,7 @@
  "admin.nav.providers": "服务商",
  "admin.nav.system": "系统",
  "admin.nav.logs": "日志",
+  "admin.nav.stats": "统计",
  "admin.users.title": "用户管理",
  "admin.users.subtitle": "查看和管理用户账户",
  "admin.users.planUpdated": "套餐已更新",
@@ -44,5 +45,105 @@
  "admin.system.waitingData": "等待数据...",
  "admin.system.purging": "清理中...",
  "admin.system.clean": "清理",
-  "admin.system.purge": "清除"
+  "admin.system.purge": "清除",
+  "admin.nav.models": "模型与订阅",
+  "admin.nav.marketing": "市场营销",
+  "admin.marketing.audience": "目标人群",
+  "admin.marketing.audienceAllUsers": "全部账户",
+  "admin.marketing.audienceInactive": "30 天未活动的账户",
+  "admin.marketing.audiencePlaceholder": "选择目标人群",
+  "admin.marketing.audiencePlanBusiness": "Business 套餐",
+  "admin.marketing.audiencePlanEnterprise": "Enterprise 套餐",
+  "admin.marketing.audiencePlanFree": "免费套餐",
+  "admin.marketing.audiencePlanPro": "Pro 套餐",
+  "admin.marketing.audiencePlanStarter": "Starter 套餐",
+  "admin.marketing.audienceWaitlist": "等候名单",
+  "admin.marketing.badgeReal": "正式",
+  "admin.marketing.badgeTest": "测试",
+  "admin.marketing.confirmSend": "将此邮件发送给 {count} 位收件人（「{audience}」）？",
+  "admin.marketing.footerNote": "发送时会自动添加包含退订链接的页脚。",
+  "admin.marketing.hidePreview": "隐藏预览",
+  "admin.marketing.history": "发送历史",
+  "admin.marketing.historyFailedSuffix": "，{count} 封失败",
+  "admin.marketing.historySent": "已发送 {sent}/{total}",
+  "admin.marketing.html": "HTML 内容",
+  "admin.marketing.htmlPlaceholder": "<html><body>…</body></html>",
+  "admin.marketing.incompleteDesc": "请填写主题和 HTML。",
+  "admin.marketing.incompleteTitle": "内容不完整",
+  "admin.marketing.loadErrorTitle": "加载错误",
+  "admin.marketing.localCheckUnavailable": "此浏览器不支持本地校验：服务器将检查事先的测试发送。",
+  "admin.marketing.networkDesc": "无法连接后端。",
+  "admin.marketing.newCampaign": "新营销活动",
+  "admin.marketing.noHistory": "暂无发送记录。",
+  "admin.marketing.preview": "预览",
+  "admin.marketing.previewEmpty": "预览：请在旁边粘贴您的 HTML。",
+  "admin.marketing.previewTitle": "邮件预览",
+  "admin.marketing.recipientsAvailable": "排除退订用户后，共有 {count} 位收件人。",
+  "admin.marketing.sendReal": "发送给 {count} 位收件人",
+  "admin.marketing.sendRefusedTitle": "发送被拒绝",
+  "admin.marketing.sendTest": "测试发送",
+  "admin.marketing.sendingDesc": "{count} 封邮件进入发送队列（间隔 0.2 秒）。退订用户已被排除。",
+  "admin.marketing.sendingTitle": "发送进行中",
+  "admin.marketing.subject": "主题",
+  "admin.marketing.subjectPlaceholder": "例如：您的翻译已完成 — 7 天内 8 折",
+  "admin.marketing.subtitle": "向目标人群发送邮件，必须先测试发送，自动附加退订链接，并保留完整历史。",
+  "admin.marketing.testDoneDesc": "邮件已送达 {email}",
+  "admin.marketing.testDoneTitle": "测试发送完成",
+  "admin.marketing.testEmail": "测试邮箱（可选）",
+  "admin.marketing.testEmailPlaceholder": "否则使用 SMTP 发件地址",
+  "admin.marketing.testFailedTitle": "测试发送失败",
+  "admin.marketing.testRequiredHint": "正式发送要求先对完全相同的内容进行测试发送（由服务器校验）。",
+  "admin.marketing.testValidated": "此内容的测试发送已通过。",
+  "admin.marketing.title": "营销 — 邮件营销",
+  "admin.marketing.unsubscribedNote": "{count} 位退订用户将被排除在所有发送之外。",
+  "admin.models.addModel": "添加模型",
+  "admin.models.businessBadge": "Business 套餐",
+  "admin.models.cancel": "取消",
+  "admin.models.catalog": "OpenRouter 目录",
+  "admin.models.catalogErrorDesc": "无法加载 OpenRouter 目录。",
+  "admin.models.catalogErrorTitle": "目录不可用",
+  "admin.models.customBadge": "定制层级",
+  "admin.models.customNote": "定制模型，与客户逐案商定。",
+  "admin.models.defaultBadge": "默认",
+  "admin.models.emptyTier": "暂无模型：将使用套餐的官方模型阵容。",
+  "admin.models.enterpriseBadge": "Enterprise 套餐",
+  "admin.models.essentialBadge": "基础 AI 层级",
+  "admin.models.essentialCost": "— 按成本系数 1 计费",
+  "admin.models.fallbackFirst": "（首选备用）",
+  "admin.models.loadErrorDesc": "HTTP {status} — 请检查管理员令牌。",
+  "admin.models.loadErrorTitle": "加载错误",
+  "admin.models.matrixDesc": "每个层级的有效模型 — 列表顺序即备用优先级。单选按钮选择默认模型，立即生效，无需重新部署。",
+  "admin.models.matrixTitle": "按套餐划分的模型矩阵",
+  "admin.models.missingConfig": "未找到 AI 层级配置。请刷新页面。",
+  "admin.models.moveDown": "下移",
+  "admin.models.moveUp": "上移",
+  "admin.models.networkDesc": "无法连接后端。",
+  "admin.models.networkTitle": "网络错误",
+  "admin.models.premiumBadge": "高级 AI 层级",
+  "admin.models.premiumCost": "— 按成本系数 5 计费",
+  "admin.models.premiumReservedNote": "仅限 Business 和 Enterprise 套餐（「openrouter_premium」引擎）。",
+  "admin.models.proBadge": "Pro 套餐",
+  "admin.models.remove": "移除",
+  "admin.models.save": "保存",
+  "admin.models.saveErrorTitle": "保存错误",
+  "admin.models.saveNetworkDesc": "无法保存配置。",
+  "admin.models.savedDesc": "新的默认模型从下一次翻译起生效，无需重新部署。",
+  "admin.models.savedTitle": "模型已保存",
+  "admin.models.saving": "保存中...",
+  "admin.models.setDefault": "将 {model} 设为默认模型",
+  "admin.models.sharedTierNote": "共享层级：Business 套餐的「openrouter」引擎同样使用基础层级 — 上面的列表对两个套餐都适用。",
+  "admin.models.subtitle": "套餐 → AI 层级矩阵。每次翻译实际使用的模型是套餐所属层级的模型：Pro 套餐绝不会触发 Premium 模型。",
+  "admin.models.title": "模型与订阅",
+  "admin.stats.aiTiers": "AI 层级（30 天）",
+  "admin.stats.includingCredits": "其中 {amount} 为积分",
+  "admin.stats.mrr": "预估 MRR",
+  "admin.stats.noTranslationsYet": "最近 30 天数据库中没有翻译记录：层级分布将随后续翻译逐步填充。",
+  "admin.stats.payments30": "30 天内 {count} 笔付款",
+  "admin.stats.refreshing": "刷新中...",
+  "admin.stats.revenue30": "收入（30 天）",
+  "admin.stats.revenueTotal": "累计已收收入",
+  "admin.stats.tiersSub": "基础 / 高级 · 经典 {classic} · 其他 {other}",
+  "admin.stats.unavailable": "业务统计不可用（{error}）。",
+  "admin.stats.waitlist": "等候名单",
+  "admin.stats.waitlistSub": "位用户等候中"
 }
--- a/frontend/src/lib/i18n/messages/zh/landing.json
+++ b/frontend/src/lib/i18n/messages/zh/landing.json
@@ -100,7 +100,7 @@
  "landing.pricing.pro.desc": "适合专业用户",
  "landing.pricing.pro.f1": "每月200份文档",
  "landing.pricing.pro.f2": "每份文档最多200页",
-  "landing.pricing.pro.f3": "AI翻译",
+  "landing.pricing.pro.f3": "基础 AI：DeepSeek V4 Flash、GLM-5.3 Flash、MiniMax M3",
  "landing.pricing.pro.f4": "包含Google",
  "landing.pricing.pro.f5": "自定义术语表和提示词",
  "landing.pricing.pro.f6": "优先支持",
@@ -109,7 +109,7 @@
  "landing.pricing.business.desc": "适合大批量需求的团队",
  "landing.pricing.business.f1": "每月1,000份文档",
  "landing.pricing.business.f2": "每份文档最多500页",
-  "landing.pricing.business.f3": "高级AI（Claude）",
+  "landing.pricing.business.f3": "高级AI（Claude 5）",
  "landing.pricing.business.f4": "所有供应商 + API",
  "landing.pricing.business.f5": "Webhook和自动化",
  "landing.pricing.business.f6": "5个团队席位",
@@ -142,5 +142,31 @@
  "landing.translate.supportedFormats": "支持 DOCX, XLSX, PPTX 或 PDF 文件",
  "landing.translate.aiAnalysis": "AI 分析中",
  "landing.translate.processing": "正在处理",
-  "landing.translate.preservingLayout": "正在保留您的排版"
+  "landing.translate.preservingLayout": "正在保留您的排版",
+  "landing.beforeAfter.seal": "相同的排版，逐字对应",
+  "landing.beforeAfter.proof1": "SmartArt 图形会重新构建，而不是被压平",
+  "landing.beforeAfter.proof2": "图表系列与坐标轴一并翻译",
+  "landing.beforeAfter.proof3": "目录按目标语言重新生成",
+  "landing.beforeAfter.targetTitle": "技术规格 — 空气处理",
+  "landing.beforeAfter.targetPara1": "通风系统必须在第二季度结束前调试运行。",
+  "landing.beforeAfter.targetPara2Before": "",
+  "landing.beforeAfter.targetTerm": "空气处理机组",
+  "landing.beforeAfter.targetPara2After": "必须按照计划达到 F7 过滤等级。",
+  "landing.beforeAfter.targetItem1": "热负荷：额定流量下 42 kW",
+  "landing.beforeAfter.targetItem2": "3 米处噪音低于 45 dB(A)",
+  "landing.formats.pill": "兼容性",
+  "landing.hero.visualCaption": "相同的排版，新的语言 — 其余一切保持不变",
+  "landing.pricing.free.name": "免费",
+  "landing.pricing.free.desc": "适合初次体验应用",
+  "landing.pricing.free.cta": "选择此套餐",
+  "landing.pricing.enterprise.name": "企业版",
+  "landing.pricing.enterprise.desc": "面向大型组织的定制解决方案",
+  "landing.pricing.enterprise.cta": "联系我们",
+  "landing.beforeAfter.sourceTitle": "Cahier des charges — Traitement d'air",
+  "landing.beforeAfter.sourcePara1": "L'installation de ventilation doit être mise en service avant la fin du deuxième trimestre.",
+  "landing.beforeAfter.sourcePara2Before": "Le",
+  "landing.beforeAfter.sourceTerm": "groupe de traitement d'air",
+  "landing.beforeAfter.sourcePara2After": "doit respecter la classe de filtration F7 conformément au planning.",
+  "landing.beforeAfter.sourceItem1": "Charge thermique : 42 kW au débit nominal",
+  "landing.beforeAfter.sourceItem2": "Niveau sonore inférieur à 45 dB(A) à 3 m"
 }
--- a/frontend/src/lib/i18n/messages/zh/pricing.json
+++ b/frontend/src/lib/i18n/messages/zh/pricing.json
@@ -41,7 +41,7 @@
  "pricing.plans.pro.feat8": "90 天历史记录",
  "pricing.plans.business.feat1": "每月 1,000 份文档",
  "pricing.plans.business.feat2": "每份文档最多 500 页",
-  "pricing.plans.business.feat3": "基础 + 高级 AI（Claude Haiku）",
+  "pricing.plans.business.feat3": "基础 + 高级 AI（Claude 5）",
  "pricing.plans.business.feat4": "所有翻译服务商",
  "pricing.plans.business.feat5": "文件最大 50 MB",
  "pricing.plans.business.feat6": "API 访问（每月 10,000 次调用）",
@@ -50,7 +50,7 @@
  "pricing.plans.business.feat9": "1 年历史记录",
  "pricing.plans.business.feat10": "高级分析",
  "pricing.plans.enterprise.feat1": "无限文档",
-  "pricing.plans.enterprise.feat2": "所有 AI 模型（GPT-5、Claude Opus 4.6…）",
+  "pricing.plans.enterprise.feat2": "所有 AI 模型（Claude 5、DeepSeek V4 Pro、GLM-5.3…）",
  "pricing.plans.enterprise.feat3": "本地部署或专属云",
  "pricing.plans.enterprise.feat4": "99.9% SLA 保障",
  "pricing.plans.enterprise.feat5": "24/7 专属支持",
@@ -108,19 +108,19 @@
  "pricing.trust.parallel.sub": "超快多线程 AI",
  "pricing.trust.availability.title": "7×24 小时可用",
  "pricing.trust.availability.sub": "99.9% 可用性保障",
-  "pricing.aiModels.title": "我们的 AI 模型 — 2026年3月",
+  "pricing.aiModels.title": "我们的 AI 模型 — 2026年9月",
  "pricing.aiModels.essential.title": "基础 AI 翻译",
  "pricing.aiModels.essential.plan": "专业版方案",
  "pricing.aiModels.essential.descPrefix": "基于",
-  "pricing.aiModels.essential.descSuffix": "— 2026 年最具性价比的 AI 模型。质量媲美前沿模型，成本仅为其一小部分。",
-  "pricing.aiModels.essential.modelName": "Essential AI模型",
-  "pricing.aiModels.essential.context": "163K 上下文 Token",
-  "pricing.aiModels.essential.value": "性价比极佳",
+  "pricing.aiModels.essential.descSuffix": "— 2026 年最具性价比的 AI 模型。",
+  "pricing.aiModels.essential.modelName": "DeepSeek V4 Flash、GLM-5.3 Flash 和 MiniMax M3",
+  "pricing.aiModels.essential.context": "上下文最多 130 万 Token（GLM-5.3 Flash）",
+  "pricing.aiModels.essential.value": "极致性价比",
  "pricing.aiModels.premium.title": "高级 AI 翻译",
  "pricing.aiModels.premium.plan": "企业版方案",
  "pricing.aiModels.premium.descPrefix": "基于",
  "pricing.aiModels.premium.descSuffix": "Anthropic 出品 — 在法律、医疗和复杂技术文档方面表现精准。",
-  "pricing.aiModels.premium.context": "200K 上下文 Token",
+  "pricing.aiModels.premium.context": "100 万上下文 Token",
  "pricing.aiModels.premium.precision": "最高精度",
  "pricing.faq.title": "常见问题",
  "pricing.faq.q1": "我可以随时更换方案吗？",
@@ -128,7 +128,7 @@
  "pricing.faq.q2": "什么是「基础 AI 翻译」？",
  "pricing.faq.a2": "这是我们的AI引擎。它能理解文档上下文，保留排版，并比传统翻译更好地处理技术术语。",
  "pricing.faq.q3": "基础 AI 和高级 AI 有什么区别？",
-  "pricing.faq.a3": "Essential AI使用优化模型（性价比极佳）。Premium AI使用Anthropic的Claude 3.5 Haiku，在法律、医学和复杂技术文档上更准确。",
+  "pricing.faq.a3": "Essential AI使用DeepSeek V4 Flash、GLM-5.3 Flash和MiniMax M3（性价比极佳）。Premium AI使用Anthropic的Claude 5，在法律、医学和复杂技术文档上更准确。",
  "pricing.faq.q4": "翻译后文档会被保留吗？",
  "pricing.faq.a4": "翻译文件根据您的方案保留（入门版 30 天、专业版 90 天、企业版 1 年）。存储和传输过程中均已加密。",
  "pricing.faq.q5": "超过月度配额会怎样？",
@@ -146,5 +146,26 @@
  "pricing.toast.paymentError": "创建支付时出错。",
  "pricing.dashboard": "仪表板",
  "pricing.okSymbol": "✓",
-  "pricing.errSymbol": "✕"
+  "pricing.errSymbol": "✕",
+  "pricing.aiModels.essential.price": "每百万 Token 低至 $0.09",
+  "pricing.aiModels.premium.modelName": "Claude 5",
+  "pricing.aiModels.premium.alternatives": "备选模型：DeepSeek V4 Pro 和 GLM-5.3",
+  "pricing.confirm.title": "确认您的订阅",
+  "pricing.confirm.subtitle": "您将被重定向到我们的安全支付服务商。",
+  "pricing.confirm.cancel": "取消",
+  "pricing.confirm.cta": "继续付款",
+  "pricing.confirm.year": "年",
+  "pricing.confirm.month": "个月",
+  "pricing.confirm.monthlyEquivalent": "按年计费 — 相当于每月 {price} 欧元。",
+  "pricing.confirm.secureNote": "可随时在个人资料中取消。付款由 Stripe 处理；您的银行卡信息绝不会经过我们的服务器。",
+  "pricing.enterprise.subject": "企业版方案咨询",
+  "pricing.enterpriseBand.cta": "联系我们",
+  "pricing.enterpriseBand.text": "大批量、专属引擎、本地部署选项 — 欢迎洽谈。",
+  "pricing.error.server": "服务器错误 {status}",
+  "pricing.freeBand.cta": "免费开始",
+  "pricing.freeBand.text": "只想先试试？免费开始 — 每月 5 份文档，无需银行卡。",
+  "pricing.header.titleBase": "总有一款套餐适合",
+  "pricing.header.titleAccent": "您的需求",
+  "pricing.toast.close": "关闭",
+  "pricing.aiModels.premium.price": "每百万 Token $2 / $10"
 }
--- a/frontend/src/lib/i18n/messages/zh/providerTheme.json
+++ b/frontend/src/lib/i18n/messages/zh/providerTheme.json
@@ -13,7 +13,7 @@
  "providerTheme.openrouter.desc": "统一访问为翻译优化的最佳开源模型。",
  "providerTheme.openrouter_premium.badge": "超级",
  "providerTheme.openrouter_premium.subBadge": "最大上下文",
-  "providerTheme.openrouter_premium.desc": "由尖端模型(GPT-4o、Claude Sonnet 4.6)辅助,适用于长文档。",
+  "providerTheme.openrouter_premium.desc": "由尖端模型(GPT-4o、Claude 5)辅助,适用于长文档。",
  "providerTheme.zai.badge": "专业",
  "providerTheme.zai.subBadge": "金融与法律",
  "providerTheme.zai.desc": "针对苛刻的业务术语(法律、金融)进行了微调的模型。",
--- a/frontend/src/lib/store.ts
+++ b/frontend/src/lib/store.ts
@@ -8,9 +8,12 @@ export const openaiModels = [
 ];

 export const openrouterModels = [
-  { id: "google/gemini-3.5-flash", name: "Gemini 3.5 Flash" },
-  { id: "deepseek/deepseek-chat", name: "DeepSeek Chat" },
-  { id: "anthropic/claude-sonnet-4.6", name: "Claude Sonnet 4.6" },
+  { id: "deepseek/deepseek-v4-flash", name: "DeepSeek V4 Flash" },
+  { id: "z-ai/glm-5.3-flash", name: "GLM-5.3 Flash" },
+  { id: "minimax/minimax-m3", name: "MiniMax M3" },
+  { id: "anthropic/claude-sonnet-5", name: "Claude 5" },
+  { id: "deepseek/deepseek-v4-pro", name: "DeepSeek V4 Pro" },
+  { id: "z-ai/glm-5.3", name: "GLM-5.3" },
 ];

 interface TranslationSettings {
@@ -101,7 +104,7 @@ const defaultSettings: TranslationSettings = {
  openaiApiKey: "",
  openaiModel: "gpt-4o-mini",
  openrouterApiKey: "",
-  openrouterModel: "deepseek/deepseek-chat-v3-0324",
+  openrouterModel: "deepseek/deepseek-v4-flash",
  libreTranslateUrl: "",
  systemPrompt: "",
  glossary: "",