.

2026-09-24 11:42:53 +02:00
parent 30918f33e7
commit ac04c674a6
5 changed files with 57 additions and 13 deletions
@@ -14,6 +14,7 @@ import { Button, CopyButton, Field, FlowerMark, Input, LoadingScreen, Picker, Sp
 import { cn } from "../lib/cn.ts";
 import { msUntilNextSolarSwitch, resolveThemeDark } from "../lib/solarTheme.ts";
 import { useT, type StringKey } from "../lib/i18n.ts";
+import { OPENAI_STATIC_VOICES } from "../lib/voice.ts";

 type Tab = "accounts" | "access" | "reports" | "models" | "stats" | "features" | "mail";

@@ -349,6 +350,8 @@ function VoicePreviewsCard() {
  const [busy, setBusy] = useState(false);
  const [msg, setMsg] = useState<{ ok: boolean; text: string } | null>(null);
  const [serverVoices, setServerVoices] = useState<{ id: string; name: string }[]>([]);
+  const [ttsProvider, setTtsProvider] = useState<string | null>(null);
+  const [ttsAvailable, setTtsAvailable] = useState(false);

  useEffect(() => {
    api.adminGetSettings()
@@ -357,7 +360,13 @@ function VoicePreviewsCard() {
        setLoaded(true);
      })
      .catch(() => setLoaded(true));
-    api.ttsInfo().then((r) => setServerVoices(r.voices)).catch(() => {});
+    api.ttsInfo()
+      .then((r) => {
+        setServerVoices(r.voices);
+        setTtsProvider(r.provider);
+        setTtsAvailable(r.available);
+      })
+      .catch(() => {});
  }, []);

  async function save(next: VoicePreview[]) {
@@ -454,8 +463,8 @@ function VoicePreviewsCard() {
                </Field>
              </span>
              <span className="w-40">
-                <Field label={t("vfeat.voice")} hint={t("vfeat.voiceHint")}>
-                  {serverVoices.length > 0 ? (
+                <Field label={t("vfeat.voice")} hint={ttsAvailable ? (ttsProvider ? `${ttsProvider} · ${t("vfeat.voiceHint")}` : t("vfeat.voiceHint")) : t("vfeat.noProvider")}>
+                  {serverVoices.length > 0 || ttsProvider !== "elevenlabs" ? (
                    <Picker
                      ariaLabel={t("vfeat.voice")}
                      value={line.voice}
@@ -465,7 +474,12 @@ function VoicePreviewsCard() {
                        if (next[i]!.text.trim() !== "") void save(next);
                      }}
                      align="left"
-                      options={[{ value: "", label: t("vfeat.voicePh") }, ...serverVoices.map((v) => ({ value: v.id, label: v.name }))]}
+                      options={[
+                        { value: "", label: t("vfeat.voicePh") },
+                        ...(serverVoices.length > 0
+                          ? serverVoices.map((v) => ({ value: v.id, label: v.name }))
+                          : OPENAI_STATIC_VOICES.map((v) => ({ value: v, label: v }))),
+                      ]}
                    />
                  ) : (
                    <Input
@@ -4,7 +4,7 @@ import { api, type UserProvider, type VoicePreview } from "../lib/api.ts";
 import { pushToast } from "../lib/toasts.ts";
 import { cn } from "../lib/cn.ts";
 import { ACCENT_PRESETS, applyAccent, currentAccent, previewAccent, type AccentState } from "../lib/accent.ts";
-import { detectLang, getVoicePrefs, listTtsVoices, pickVoice, playText, setVoicePrefs, stopPlayback, ttsSupported, type VoicePrefs } from "../lib/voice.ts";
+import { detectLang, getVoicePrefs, listTtsVoices, OPENAI_STATIC_VOICES, pickVoice, playText, setVoicePrefs, stopPlayback, ttsSupported, type VoicePrefs } from "../lib/voice.ts";
 import { FEATURES, setFeature, useFeatures } from "../lib/features.ts";
 import type { Conversation, FilePreview, User } from "../lib/types.ts";
 import { useT, type StringKey } from "../lib/i18n.ts";
@@ -978,8 +978,11 @@ function VoiceSection() {
        />
      </Field>
      {(prefs.engine !== "browser") && (
-        ttsServer?.available ? (
-          ttsServer.voices.length > 0 ? (
+        <>
+          {!ttsServer?.available && (
+            <p className="rounded-2xl bg-amber-500/10 px-4 py-2.5 text-sm text-amber-700 dark:text-amber-400">{t("voice.noServer")}</p>
+          )}
+          {ttsServer && ttsServer.voices.length > 0 ? (
            <Field label={t("voice.serverVoice")}>
              <Picker
                ariaLabel={t("voice.serverVoice")}
@@ -990,7 +993,7 @@ function VoiceSection() {
                options={[{ value: "", label: t("voice.defaultVoice") }, ...ttsServer.voices.map((v) => ({ value: v.id, label: v.name }))]}
              />
            </Field>
-          ) : ttsServer.provider === "elevenlabs" ? (
+          ) : ttsServer?.provider === "elevenlabs" ? (
            <Field label={t("voice.serverVoice")} hint={t("voice.serverVoiceHint")}>
              <Input
                value={prefs.serverVoice ?? ""}
@@ -999,12 +1002,21 @@ function VoiceSection() {
                maxLength={64}
              />
            </Field>
-          ) : (
+          ) : ttsServer?.provider === "piper" ? (
            <p className="rounded-2xl bg-stone-100 px-4 py-2.5 text-sm opacity-70 dark:bg-zinc-800">{t("voice.piperNote")}</p>
-          )
-        ) : (
-          <p className="rounded-2xl bg-amber-500/10 px-4 py-2.5 text-sm text-amber-700 dark:text-amber-400">{t("voice.noServer")}</p>
-        )
+          ) : (
+            <Field label={t("voice.serverVoice")} hint={t("vfeat.noProvider")}>
+              <Picker
+                ariaLabel={t("voice.serverVoice")}
+                icon={Volume2}
+                value={prefs.serverVoice ?? ""}
+                onChange={(v) => update({ serverVoice: v || null })}
+                align="left"
+                options={[{ value: "", label: t("voice.defaultVoice") }, ...OPENAI_STATIC_VOICES.map((v) => ({ value: v, label: v }))]}
+              />
+            </Field>
+          )}
+        </>
      )}
      <Field label={t("voice.voice")}>
        <Picker
@@ -608,6 +608,8 @@ const en = {
  "vfeat.voice": "Server voice",
  "vfeat.voiceHint": "ElevenLabs id / OpenAI name, empty = default",
  "vfeat.voicePh": "auto",
+  "vfeat.noProvider": "No neural provider yet — add ELEVENLABS_API_KEY or OPENAI_API_KEY to activate.",
+
  "center.ttsTitle": "Neural voice playback",
  "center.ttsDesc": "Server text-to-speech for message playback (ElevenLabs / OpenAI / piper). Off = browser speech only.",
  "voice.engine": "Playback engine",
@@ -1316,6 +1318,8 @@ const es: Dict = {
  "vfeat.voice": "Server voice",
  "vfeat.voiceHint": "ElevenLabs id / OpenAI name, empty = default",
  "vfeat.voicePh": "auto",
+  "vfeat.noProvider": "No neural provider yet — add ELEVENLABS_API_KEY or OPENAI_API_KEY to activate.",
+
  "center.ttsTitle": "Neural voice playback",
  "center.ttsDesc": "Server text-to-speech for message playback (ElevenLabs / OpenAI / piper). Off = browser speech only.",
  "voice.engine": "Playback engine",
@@ -2019,6 +2023,8 @@ const fr: Dict = {
  "vfeat.voice": "Server voice",
  "vfeat.voiceHint": "ElevenLabs id / OpenAI name, empty = default",
  "vfeat.voicePh": "auto",
+  "vfeat.noProvider": "No neural provider yet — add ELEVENLABS_API_KEY or OPENAI_API_KEY to activate.",
+
  "center.ttsTitle": "Neural voice playback",
  "center.ttsDesc": "Server text-to-speech for message playback (ElevenLabs / OpenAI / piper). Off = browser speech only.",
  "voice.engine": "Playback engine",
@@ -2710,6 +2716,8 @@ const ru: Dict = {
  "vfeat.voice": "Server voice",
  "vfeat.voiceHint": "ElevenLabs id / OpenAI name, empty = default",
  "vfeat.voicePh": "auto",
+  "vfeat.noProvider": "No neural provider yet — add ELEVENLABS_API_KEY or OPENAI_API_KEY to activate.",
+
  "center.ttsTitle": "Neural voice playback",
  "center.ttsDesc": "Server text-to-speech for message playback (ElevenLabs / OpenAI / piper). Off = browser speech only.",
  "voice.engine": "Playback engine",
@@ -3406,6 +3414,8 @@ const it: Dict = {
  "vfeat.voice": "Server voice",
  "vfeat.voiceHint": "ElevenLabs id / OpenAI name, empty = default",
  "vfeat.voicePh": "auto",
+  "vfeat.noProvider": "No neural provider yet — add ELEVENLABS_API_KEY or OPENAI_API_KEY to activate.",
+
  "center.ttsTitle": "Neural voice playback",
  "center.ttsDesc": "Server text-to-speech for message playback (ElevenLabs / OpenAI / piper). Off = browser speech only.",
  "voice.engine": "Playback engine",
@@ -4093,6 +4103,8 @@ const pt: Dict = {
  "vfeat.voice": "Server voice",
  "vfeat.voiceHint": "ElevenLabs id / OpenAI name, empty = default",
  "vfeat.voicePh": "auto",
+  "vfeat.noProvider": "No neural provider yet — add ELEVENLABS_API_KEY or OPENAI_API_KEY to activate.",
+
  "center.ttsTitle": "Neural voice playback",
  "center.ttsDesc": "Server text-to-speech for message playback (ElevenLabs / OpenAI / piper). Off = browser speech only.",
  "voice.engine": "Playback engine",
@@ -90,6 +90,9 @@ export interface SpeakOpts {

 export type VoiceTimbre = "masculine" | "feminine" | "any";

+/** Stock OpenAI voices (static catalog — selectable even before any key). */
+export const OPENAI_STATIC_VOICES = ["alloy", "echo", "fable", "onyx", "nova", "shimmer"];
+
 // Name fragments hinting a masculine / feminine voice (Chrome, Edge,
 // Safari and Android voices carry such markers; absence falls back to
 // any same-language voice — never an error).