docs(readme): nota sui limiti delle voci TTS imposti dal server (v0.3.2)

Le voci selezionabili per la lettura ad alta voce dipendono dall'allowlist
Edge TTS del server hermes-webui: se il server rifiuta una voce (400
invalid voice) si sceglie tra quelle ammesse, si lascia vuoto (default
del server) o si estende l'allowlist sul proprio server. Solo
documentazione: nessuna modifica al comportamento dell'app.

README.md

@@ -40,6 +40,13 @@ requires a server address.
 | Auto-stop pause / Mic sensitivity | voice-dialog tuning (1.5 s / medium) |
 | Max recording time | voice-dialog safety cap (60 s) |
 
+> **Read-aloud voices depend on your server.** The selected voice is sent to
+> `hermes-webui`'s `/api/tts`, which only accepts voices present in its Edge
+> TTS allowlist (`api/routes.py`; upstream ships Chinese, English, French and
+> Indonesian voices). If your server rejects a voice (`400 invalid voice`),
+> pick another one, leave the voice empty (server default), or add your
+> language's voices to your own `hermes-webui`.
+
 ## Build
 
 See [docs/BUILD.md](docs/BUILD.md) (Sailfish SDK, `sfdk build` / `sfdk deploy`)
@@ -48,6 +55,6 @@ speaks (validated live against a real instance).
 
 ## Status
 
-Version 0.3.1 — protocol and C++ core validated against a real `hermes-webui`
+Version 0.3.2 — protocol and C++ core validated against a real `hermes-webui`
 (login, sessions, streaming chat, TTS, STT, voice dialog); on-device testing
 still ongoing.