Fish Audio Docs
Integration GuidesThird-Party Integrations
Back to site

Open Notebook Setup

Add Fish Audio OpenAI-compatible TTS credentials and a speech model to Open Notebook for podcasts and audio summaries.

Connect Fish Audio TTS to Open Notebook

Open Notebook supports an OpenAI-compatible TTS provider for podcast generation, with separate credentials, model, and voice settings.

Add credentials

Open:

Settings → API Keys → Add Credential → OpenAI-Compatible

Enter:

SettingValue
TTS Base URLhttps://fishaudio.org/v1
API KeyYour Fish Audio API key

Save and run Test Connection. Then open Settings → Models → Text-to-Speech and add:

SettingValue
Provideropenai_compatible
Model Namefishaudio-s21pro-flash
Display NameFish Audio S2.1 Pro

Generate a podcast

Select the new model in the podcast or audio generator and set Voice to the system test Voice ID. Replace it with your preferred voice after the first successful generation.

Troubleshooting

  • Docker cannot connect: Fish Audio uses a public HTTPS URL; do not replace it with host.docker.internal.
  • Fish Audio is missing from models: select openai_compatible as the provider.
  • Default voice fails: manually enter a Fish Audio Voice ID.
  • 401/402: verify credentials and API credits.

Dify Setup

Use Fish Audio TTS with the Dify Podcast Studio plugin to generate single-speaker or two-host podcast audio.

LiveKit Agents Setup

Configure the Fish Audio Base URL, model, and voice in the LiveKit Agents OpenAI TTS plugin.

On this page

Connect Fish Audio TTS to Open NotebookAdd credentialsGenerate a podcastTroubleshootingRelated documentation