Text-To-Speech Agent with Fish Audio

This Ableton Live extension that turns typed text or a dropped .txt file into generated audio using Fish Audio(https://fish.audio/).

What You Need

• Ableton Live with extension support.
• The audio-synthsis-agent-live-extension.ablx file.
• A Fish Audio API key.
• A Fish Audio reference_id for the voice ID you want to use.

Install the .ablx File

1. Open Ableton Live 12 Beta.
2. Find the .ablx file in Finder.
3. Open Settings in ableton and go to Extensions. Disable the Developer Mode at the buttom, then drag the .ablx file into the Ableton Live window.
4. Restart Ableton Live if Ableton asks you to.

Open the Extension

1. Go to Session View in Ableton Live.

2. Right-click an audio track or an empty clip slot.

3. Choose the Text to Audio Agent action:

   • On an audio track: Open audio-synthsis-agent-live-extension
   • On a clip slot: Generate audio here

Generate Audio

1. Enter text in the Text box, or drag a .txt file onto Drop a .txt file here.

2. Enter your Fish Audio reference_id.

3. Enter your Fish Audio API key.

4. Leave the default settings unless you need to change them:

   • Endpoint: Fish Audio API endpoint.
   • Model: Fish Audio model.
   • Format: output file format, usually wav.
   • Sample rate: usually 44100.
   • Output directory: temporary folder for generated files.

5. Adjust optional voice controls if needed:

   • Temperature
   • Top P
   • Speed
   • Output filename

6. Click Generate Audio.

Where the Clip Appears

If you opened the extension from an audio track:

• The generated clip goes into the first empty Session View clip slot on that track.
• If every slot is full, the extension replaces the oldest clip previously generated by Text to Audio Agent.
• It does not replace unrelated user clips if no older Text to Audio Agent clip exists.

If you opened the extension from a clip slot:

• The generated clip goes into that exact slot.
• If the slot already has a clip, that clip is replaced.

Loading .txt Files

The extension uses drag and drop only.

1. Drag a .txt file from Finder.
2. Drop it onto the dialog area labeled Drop a .txt file here.
3. The text will appear in the Text box.

Troubleshooting

If generation fails:

• Check that your Fish Audio API key is correct.
• Check that your reference_id is correct.
• Check that your internet connection is working.
• Try a shorter text prompt.
• Make sure the selected Ableton target is an audio track or clip slot.

If Ableton says the Extension Host stopped running, check the extension log:

text-to-audio-agent.log

The log is written in Ableton's extension storage folder when available.