lb-pocket-tts-skill
par leonaaardob
Media & StreamingRésumé
Generate speech from text using Kyutai Pocket TTS - lightweight, CPU-friendly, streaming TTS with voice cloning.
Description
Ce que fait ce skill
Pocket-TTS generates speech from text using the Kyutai model locally on CPU, with no GPU required. It supports voice cloning from short audio samples and streaming output with under 200ms latency. English only, approximately 6x faster than real-time on modern hardware.
Runs entirely on CPU without cloud API dependencies, making it faster to set up and cheaper to operate than hosted TTS services for local or offline workflows.
Quand l'utiliser
- Narrating documentation without a cloud TTS subscription
- Cloning a custom voice for consistent audio branding
- Generating speech offline on a laptop during travel
- Prototyping voice interfaces without API rate limits
- Batch-converting written content to audio files
Exemple de workflow
User asks: generate speech from a script using my own voice sample
- 1
Export voice embedding from audio sample with pocket-tts export-voice voice.wav voice.safetensors
- 2
Load the TTSModel and retrieve voice state from the safetensors file
- 3
Call model.generate_audio with the voice state and input text
- 4
Save the returned audio array as a 24kHz WAV file using scipy
A 24kHz mono WAV file with speech in the cloned voice
Prérequis
- HuggingFace account (to access kyutai/tts-voices or kyutai/pocket-tts model weights)
Installation
clawhub install leonaaardob/lb-pocket-tts-skillExécute la commande CLI, ou colle l'instruction dans Claude Code, Cursor, Codex ou l'agent de code que tu utilises.