VoiceLab Documentation

Add LLM completions, speech generation, transcription, noise reduction, and realtime audio to your product. Use HTTP APIs, WebSockets, or the official TTS and STT SDKs.

How VoiceLab works

Use the REST API to generate LLM completions or speech, transcribe files, or remove background noise with Voice Isolator. Use authenticated WebSockets when audio must move in realtime. Developer requests use an API key. LLM completions also support SSE and return X-LLM-Request-ID for usage lookup.

Choose your path

• Start here

  Quickstart

  Create a key and send your first request.

• SDKs

  Python and TypeScript

  Install an SDK and generate audio.

• Reference

  API reference

  Open endpoints, payloads, and responses.

Build with VoiceLab

• LLM

  Generate text and function calls with JSON or streaming responses.

• Text to Speech

  Generate mono 24 kHz WAV audio.

• Speech to Text

  Transcribe audio with timing metadata.

• Voice Isolator

  Remove background noise, optionally restore speech, and export audio.

• Realtime audio

  Stream audio over WebSockets.

Core resources

• Authentication and API keys
• Errors, limits, and retries
• Documentation for AI agents