Elevenlabs Audio Generator

Audio Generator (ElevenLabs)

Triggers: Human says "generate audio", "create music", "sound effects", "voice clip", "elevenlabs", "audio generator", "text to speech", "make a sound", "create audio".

Generate music, sound effects, and voice clips using the ElevenLabs API.

Tool Mapping

Setup

ElevenLabs Credentials

Call `integration_get_credentials` via the Tools API at http://host.docker.internal:3000/v1/tools/list
with parameters: { integrationId: "elevenlabs" }

Extract the api_key value. If not configured, tell the user to set up the ElevenLabs integration first.

Default Workflow

1. Check credentials

Verify ElevenLabs API key is available by calling integration_get_credentials via the Tools API.

2. Create the project

create_project({
  project_name: "<project-name>",
  content: "---\ntitle: <Project Name>\nstatus: active\ntags: [audio, elevenlabs]\n---\n# <Project Name>\n\n<description of what audio to generate>"
})

3. Create output directories

fs_mkdir({ path: "~/sulla/projects/<project-name>/audio/music" })
fs_mkdir({ path: "~/sulla/projects/<project-name>/audio/sfx" })
fs_mkdir({ path: "~/sulla/projects/<project-name>/audio/voice" })

4. Generate audio

Music Generation

curl -X POST "https://api.elevenlabs.io/v1/music/generate" \
  -H "xi-api-key: <API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "<descriptive music prompt>",
    "duration_seconds": 30,
    "output_format": "mp3_44100_128"
  }' \
  --output ~/sulla/projects/<project-name>/audio/music/<filename>.mp3

Sound Effects

curl -X POST "https://api.elevenlabs.io/v1/sound-effects/generate" \
  -H "xi-api-key: <API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "<descriptive SFX prompt>",
    "duration_seconds": 5,
    "output_format": "mp3_44100_128"
  }' \
  --output ~/sulla/projects/<project-name>/audio/sfx/<filename>.mp3

Voice / Text-to-Speech

curl -X POST "https://api.elevenlabs.io/v1/text-to-speech/<voice_id>" \
  -H "xi-api-key: <API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "<text to speak>",
    "model_id": "eleven_multilingual_v2",
    "voice_settings": {
      "stability": 0.5,
      "similarity_boost": 0.75
    }
  }' \
  --output ~/sulla/projects/<project-name>/audio/voice/<filename>.mp3

5. Present to user

After generating, present each audio file as a clickable link:

[🎵 Listen: background-music.mp3](file:///Users/<user>/sulla/projects/<project-name>/audio/music/background-music.mp3)

Directory Structure

~/sulla/projects/<project-name>/
├── PROJECT.md
├── audio/
│   ├── music/          # Generated music tracks
│   ├── sfx/            # Generated sound effects
│   └── voice/          # Generated voice clips
└── README.md

Tips

• Music prompts: Be descriptive about genre, tempo, mood, instruments. E.g., "upbeat electronic lo-fi beat with soft piano and vinyl crackle, 90 BPM"
• SFX prompts: Describe the sound precisely. E.g., "mechanical keyboard typing rapidly, close-up recording"
• Voice: Use eleven_multilingual_v2 model for best quality. Default voice IDs can be listed via the ElevenLabs API.
• Formats: mp3_44100_128 is good default. Other options: mp3_44100_192, pcm_16000, pcm_44100