kokoro-tts
par edkief
Speech & TranscriptionRésumé
Generate spoken audio from text using the local Kokoro TTS engine.
Description
Ce que fait ce skill
Converts text to spoken audio using a local or remote Kokoro TTS engine. Outputs an MP3 file that OpenClaw picks up and sends as an audio attachment. Voice, speed, and API endpoint are configurable.
Runs locally by default, so no cloud API key or per-request cost is needed.
Quand l'utiliser
- Sending a voice message reply in a chat
- Reading out a summary or report aloud
- Generating audio for a notification or alert
- Creating a spoken version of a drafted email
- Testing different TTS voices for a project
Exemple de workflow
User asks: "Say 'Your build is ready' in a professional voice"
- 1
Identifies the af_nova voice as appropriate for a professional tone
- 2
Runs tts.js with the text and af_nova voice ID
- 3
Script calls the Kokoro API at KOKORO_API_URL
- 4
Receives the generated MP3 and outputs the MEDIA: path
MEDIA: media/tts_1706745000000.mp3 — OpenClaw sends the file as an audio attachment
Prérequis
- Kokoro TTS server running locally or remotely
- KOKORO_API_KEY environment variable (if your Kokoro instance requires auth)
Installation
clawhub install edkief/kokoro-ttsExécute la commande CLI, ou colle l'instruction dans Claude Code, Cursor, Codex ou l'agent de code que tu utilises.