Back to Skills
    🦞

    kokoro-tts

    Generate spoken audio from text using the local Kokoro TTS engine.

    By @edkief
    View on GitHub
    SKILL.md
    ---
    name: kokoro-tts
    description: Generate spoken audio from text using the local Kokoro TTS engine. Use when the user asks to "say" something, requests a voice message, or wants text converted to speech.
    ---
    
    # Kokoro TTS
    
    This skill allows you to generate high-quality AI speech using a local or remote Kokoro-TTS instance.
    
    ## Configuration
    
    The skill uses the `KOKORO_API_URL` environment variable to locate the API.
    
    - **Default:** `http://localhost:8880/v1/audio/speech`
    - **To Configure:** Add `KOKORO_API_URL=http://your-server:port/v1/audio/speech` to your `.env` file or environment.
    
    ## Usage
    
    To generate speech, run the included Node.js script.
    
    ### Command
    
    ```bash
    node skills/kokoro-tts/scripts/tts.js "<text>" [voice] [speed]
    ```
    
    - **text**: The text to speak. Wrap in quotes.
    - **voice**: (Optional) The voice ID. Defaults to `af_heart`.
    - **speed**: (Optional) Speech speed (0.25 to 4.0). Defaults to `1.0`.
    
    ### Example
    
    ```bash
    node skills/kokoro-tts/scripts/tts.js "Hello Ed, this is Theosaurus speaking." af_nova
    ```
    
    ### Output
    
    The script will output a single line starting with `MEDIA:` followed by the path to the generated MP3 file. OpenClaw will automatically pick this up and send it as an audio attachment.
    
    Example Output:
    `MEDIA: media/tts_1706745000000.mp3`
    
    ## Available Voices
    
    Common choices:
    - `af_heart` (Default, Female, Warm)
    - `af_nova` (Female, Professional)
    - `am_adam` (Male, Deep)
    - `bf_alice` (British Female)
    
    For a full list, see [references/voices.md](references/voices.md) or query the API.