koelab is a local voice conversion and TTS workbench — “Ollama for Voice” — unifying voice conversion, Japanese TTS, multilingual TTS, WebUI workflows, and real-time CLI workflows in one project.

Key Features:

Quick Start:

pip install koelab

# Pull a voice model
koelab pull seed-vc

# Convert voice
koelab run seed-vc --input source.wav --reference target_voice.wav --output converted.wav

Core Commands:

Command Description
koelab pull <model> Download a voice model
koelab run <model> Run voice conversion
koelab speak <model> Synthesize speech from text
koelab live <model> Real-time voice conversion from mic
koelab serve Start REST API server + WebUI
koelab doctor Inspect runtime dependencies and setup problems
koelab models Show all available models from registry

Technologies:

Use Cases:

For developers and creators who want a local, hackable alternative to hosted voice services. Ideal for voice conversion experiments, multilingual TTS workflows, real-time voice transformation, and privacy-conscious audio processing.

GitHub Repository