refactor(config): niente configurazione cablata — tutto parametrizzato (v0.3.0)

Prima di pubblicare: eliminati dai sorgenti tutti i valori "cablati".

- settings: default neutri (URL, voce, engine vuoti = "decide il server");
  nuove chiavi ttsVoices/ttsEngines (preset dei selettori) e vadMaxMs
  (durata massima registrazione, prima fissa a 60 s nel codice)
- ApiClient: guard su baseUrl vuoto (checkAuth/login si comportano senza
  rete), voce/engine omessi nella richiesta TTS se vuoti, Content-Type
  audio coerente col formato reale del file (ogg/wav/mp3/m4a/opus)
- SettingsPage: selettori voce/engine costruiti dai preset configurabili
  (Repeater dentro ContextMenu = pattern provato su tide) + campi per
  modificare i preset; LoginPage richiede indirizzo server non vuoto
- VoiceRecorder: tetto di registrazione da configurazione
- README: sezione "Configuration" completa; docs neutralizzate (niente
  URL/valori personali: repo public-friendly)

README.md

@@ -9,18 +9,37 @@ chat with your Hermes agent from the phone — and use it with your **voice**.
 - Sessions list, new session, open any session
 - Chat with **live streaming** replies (SSE token deltas)
 - **Voice dictation**: tap Mic, speak, tap Stop → server-side Whisper
-  transcription lands in the composer
-- **Read aloud**: server-side TTS (edge engine, Italian voices) played through
+  transcription lands in the composer (optionally sent right away)
+- **Read aloud**: server-side TTS (voice/engine configurable) played through
   QMediaPlayer; optional auto-read of every completed reply
+- **Voice dialog (hands-free)**: continuous listen → transcribe → send →
+  read-aloud loop, auto-stopping on speech pauses (tunable)
 - Cancel a running turn, resync from server state, cover page
 
 ## Requirements
 
 - Sailfish OS 4.4+ (developed against Sailfish 5.1)
-- A running `hermes-webui` instance reachable over HTTPS (default:
-  `https://hermes.hackatoniclife.com`)
+- A running `hermes-webui` instance reachable over HTTPS (the address is
+  configured in the app — nothing is compiled in)
 - Sailjail permissions `Internet;Audio;Microphone` (declared in the .desktop)
 
+## Configuration
+
+No configuration is compiled into the binary: everything lives in the app's
+Settings (persisted in `~/.config/harbour/hermes.conf`, sandbox-persistent).
+Defaults are neutral — empty means "the server decides" — and the login screen
+requires a server address.
+
+| Setting | Meaning |
+|---|---|
+| Server | Hermes Web UI address, e.g. `https://hermes.example.com` (required) |
+| Read-aloud voice / TTS engine | empty = server defaults |
+| Voice/engine presets | comma-separated lists used to fill the pickers |
+| Read replies aloud | auto-play every reply (off by default) |
+| Send right after dictation | auto-send when dictation lands (on by default) |
+| Auto-stop pause / Mic sensitivity | voice-dialog tuning (1.5 s / medium) |
+| Max recording time | voice-dialog safety cap (60 s) |
+
 ## Build
 
 See [docs/BUILD.md](docs/BUILD.md) (Sailfish SDK, `sfdk build` / `sfdk deploy`)
@@ -29,6 +48,6 @@ speaks (validated live against a real instance).
 
 ## Status
 
-Version 0.2.1 — protocol and C++ core validated against a real `hermes-webui`
-(login, sessions, streaming chat, TTS, STT); on-device testing (audio paths,
-Sailjail prompts, keyboard handling) still pending.
+Version 0.3.0 — protocol and C++ core validated against a real `hermes-webui`
+(login, sessions, streaming chat, TTS, STT, voice dialog); on-device testing
+still ongoing.