voice

voice — ɳSelf documentation.

Voice Plugin

Speech-to-text, text-to-speech, and voicebot with multi-provider support. Pro plugin.

Requires: Basic license tier or higher. nself license set nself_pro_...

Install

nself license set nself_pro_xxxxx...
nself plugin install voice

What 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 records
np_voice_transcripts: STT transcription results
np_voice_recordings: stored audio references

Nginx Routes

Route	Target
`/voice/`	Voice processing API