Fish Tts

What This Skill Does

The Fish TTS skill provides a robust interface for the Fish Audio S1 text-to-speech engine, enabling OpenClaw agents to generate lifelike audio from text inputs. Beyond simple synthesis, this integration allows for the seamless streaming or saving of audio files directly to a user-configured NextCloud instance via WebDAV. It serves as a bridge between high-quality AI speech generation and cloud-based file management, making it an ideal tool for users who want to create narrated content, audio logs, or accessibility features directly within their OpenClaw-enabled environment.

Installation

To integrate this skill, ensure your system has access to the OpenClaw CLI. Execute the following command in your terminal: clawhub install openclaw/skills/skills/gtank1/fish-tts. Once installed, you must define the necessary environment variables in your local configuration, specifically NEXTCLOUD_USER, NEXTCLOUD_PASS, NEXTCLOUD_URL, and FISH_AUDIO_S1_URL. Ensure the Fish Audio S1 server is reachable at the specified local address. Failure to configure these variables will result in failed file uploads to your cloud storage.

Use Cases

This skill is highly versatile for creative and professional workflows. You can use it to convert long-form articles or documents into high-quality podcasts for on-the-go listening. Additionally, it is perfect for developers building automated notification systems where text-based alerts need to be converted into spoken audio files. For content creators, it acts as a rapid prototyping engine for voiceovers, allowing for the quick generation of audio samples using professional or emotional voice profiles.

Example Prompts

1. "Use the voice em_michael to narrate the following text and upload it to my NextCloud folder: Welcome to the daily briefing."
2. "Generate a happy sounding clip using af_bella that says: I am so excited to announce the new project launch!"
3. "Convert this news report to a professional audio file using em_pierre and save it to the /Openclaw directory on my server."

Tips & Limitations

To get the best results, use emotion tags like [happy], [sad], or [excited] to modify the prosody of the output. While the engine supports various global voices, it performs optimally in English. For testing purposes, prioritize standard quality settings to save compute resources, as high-quality generation is more intensive. Ensure your NextCloud directory exists before running the upload command to avoid 404 path errors. If you experience latency, verify your local network connection to the Fish Audio S1 server.