Python SDK

The official Python SDK for Lokutor: text-to-speech, transcription and full-duplex voice agents over the same API as the JavaScript SDK, with the same class names.

What it covers

TTSClient: stream speech from text, play it locally or collect the PCM.
STTClient and SpeechToTextClient: transcribe a recorded clip, or a live microphone.
VoiceAgentClient: a conversation with an agent through your microphone and speakers, with tool calling.
Languages: speech in English, Spanish, Catalan, Galician, Basque, Portuguese, French, Italian and German; recognition in English, Spanish, Portuguese, French, Italian and German.
lokutor.audio_utils: PCM/float conversion, resampling (with anti-aliasing), streaming resampling and RMS helpers.

Installation

Python 3.8 or later.

pip install lokutor

That is enough to call the API: synthesize to bytes, transcribe files. To play or record audio on the local machine (play=True, start_conversation(), SpeechToTextClient.start()), add the audio extra. It needs the PortAudio library first:

brew install portaudio            # macOS
sudo apt install portaudio19-dev  # Debian / Ubuntu
pip install "lokutor[audio]"

Python Quickstart
Python API Reference
Examples: python/examples in the SDK repository (simple_tts.py, simple_chat.py, advanced_chat.py)

Get Started

SDK

API Reference

Resources

What it covers

Installation

Next

​What it covers

​Installation

​Next

What it covers

Installation

Next