Update desktop app to android app capabilities.

.env.example

@@ -21,15 +21,14 @@ DESK_API_KEY=
 #WAKE_WORD_THRESHOLD=0.5
 #WAKE_CHECK_INTERVAL_SECONDS=1.2
 
-# ── STT (Deepgram) ──────────────────────────────────────────────────────────
-DEEPGRAM_API_KEY=
-#DEEPGRAM_MODEL=nova-3
-
-# ── TTS (ElevenLabs) — omit to use offline TTS only ─────────────────────────
-ELEVENLABS_API_KEY=
+# ── STT / TTS (server-hosted — same /desk/stt and /desk/tts the Android app
+# uses) ──────────────────────────────────────────────────────────────────────
+# No Deepgram or ElevenLabs account needed here: both go through
+# BOLT_SERVER_URL/DESK_API_KEY above, the same as the rest of this file.
+# ELEVENLABS_VOICE_ID just tells the server which voice to request — omit it
+# to fall back to offline pyttsx3 TTS instead (STT still works either way).
 ELEVENLABS_VOICE_ID=
-#ELEVENLABS_MODEL_ID=eleven_flash_v2
-#TTS_SAMPLE_RATE=24000
+#TTS_SAMPLE_RATE=16000
 
 # Ask Bolt to use a different voice (or another language) and the server
 # picks one from the ElevenLabs voice library and tags the reply with it.
@@ -40,10 +39,14 @@ ELEVENLABS_VOICE_ID=
 #VOICE_STICKY=true
 
 # ── Multi-voice dialogue (ElevenLabs Text to Dialogue) ──────────────────────
+# The one feature the server has no endpoint for, so this is the only place
+# in the whole app that still needs a local ElevenLabs API key — everything
+# else (the normal reply voice, transcription) goes through the server above.
 # Lets Bolt play a short scene in several voices with delivery tags the v3
 # model acts on ("[cheerfully] Hello", "[whispering] He is lying"), driven by
 # the server through a relayed `dialoguectl` command. Name the cast here —
 # "self" always means whatever voice the pet is currently using.
+ELEVENLABS_API_KEY=
 #DIALOGUE=true
 #DIALOGUE_MODEL_ID=eleven_v3
 #DIALOGUE_VOICES=narrator:9BWtsMINqrJLrRacOk9x,villain:IKne3meq5aSn9XLyUdCD
@@ -57,9 +60,6 @@ ELEVENLABS_VOICE_ID=
 #SELF_RESTART=true
 #SELF_RESTART_MAX=5
 #SELF_RESTART_WINDOW_SECONDS=900
-# Used instead of ELEVENLABS_MODEL_ID whenever the server picked the voice
-# or the reply has non-ASCII in it — the flash_v2 default is English-only.
-#ELEVENLABS_MULTILINGUAL_MODEL_ID=eleven_flash_v2_5
 
 # ── Audio devices (optional — leave blank for the system default) ──────────
 #MIC_DEVICE=
@@ -144,8 +144,9 @@ ELEVENLABS_VOICE_ID=
 # ── Latency: streaming the reply and the transcript ─────────────────────────
 # STREAMING_REPLIES speaks each sentence as the server generates it, instead of
 # waiting out the whole model call before the first word. STT_STREAMING sends
-# mic frames to Deepgram as you talk, so the transcript is ready the moment you
-# stop. Both fall back to the old path automatically if anything goes wrong.
+# mic frames to the server's /desk/stt relay as you talk, so the transcript is
+# ready the moment you stop, instead of uploading the whole clip afterward.
+# Both fall back to the old path automatically if anything goes wrong.
 # VAD_SILENCE_END_SEC is the other half: it is dead air on every single turn,
 # so 0.8-1.0 feels markedly snappier than the 1.2 default.
 #STREAMING_REPLIES=true