voice
voice — ɳSelf documentation.
Voice Plugin
Speech-to-text, text-to-speech, and voicebot with multi-provider support. Max plugin.
Requires: A paid bundle license or ɳSelf+.
nself license set nself_pro_...
Install
nself license set nself_pro_xxxxx...
nself plugin install voiceWhat It Does
Provides a unified REST API for voice processing: transcribes audio to text (STT via Whisper and other providers) and converts text to speech (TTS). Includes a voicebot mode that chains STT → AI reasoning → TTS for voice-interactive applications. Records and stores voice sessions.
Configuration
| Env Var | Default | Description |
|---|---|---|
VOICE_PORT |
3714 |
Voice service port |
VOICE_STT_PROVIDER |
whisper |
STT provider: whisper, deepgram, assembly |
VOICE_TTS_PROVIDER |
openai |
TTS provider: openai, elevenlabs, azure |
OPENAI_API_KEY |
none | Required for OpenAI STT/TTS |
DEEPGRAM_API_KEY |
none | Deepgram API key (optional) |
ELEVENLABS_API_KEY |
none | ElevenLabs API key (optional) |
VOICE_RECORDING_ENABLED |
true |
Store audio recordings |
Ports
| Port | Purpose |
|---|---|
| 3714 | Voice service REST API |
Database Tables
3 tables added to your Postgres database:
np_voice_sessions: voice session recordsnp_voice_transcripts: STT transcription resultsnp_voice_recordings: stored audio references
Nginx Routes
| Route | Target |
|---|---|
/voice/ |
Voice processing API |