Agent skills for ElevenLabs developer products. These skills follow the Agent Skills specification and can be used with any compatible AI coding assistant.
npx skills add elevenlabs/skills| Skill | Description |
|---|---|
| text-to-speech | Convert text to lifelike speech using ElevenLabs' AI voices |
| speech-to-text | Transcribe audio files to text with timestamps |
| agents | Build conversational voice AI agents |
| sound-effects | Generate sound effects from text descriptions |
| music | Generate music tracks using AI composition |
| voice-isolator | Remove background noise and isolate vocals/speech from audio |
| setup-api-key | Guide through obtaining and configuring an ElevenLabs API key |
All skills require an ElevenLabs API key. Set it as an environment variable:
export ELEVENLABS_API_KEY="your-api-key"Get your API key from the setup-api-key skill or use the ElevenLabs dashboard.
Most skills include examples for:
- Python -
pip install elevenlabs - JavaScript/TypeScript -
npm install @elevenlabs/elevenlabs-js - cURL - Direct REST API calls
JavaScript SDK Warning: Always use
@elevenlabs/elevenlabs-js. Do not usenpm install elevenlabs(that's an outdated v1.x package).
See the installation guide in any skill's references/ folder for complete setup instructions including migration from deprecated packages.
The evals/ directory contains trigger and functional evaluations for all skills.
# Run all evaluations (trigger + functional)
python3 evals/run_all.py -v
# Trigger evals only — tests if skills fire for the right queries (~3 min)
python3 evals/run_all.py --trigger-only -v
# Functional evals only — tests if skills produce correct output (~15 min)
python3 evals/run_all.py --functional-only -v
# Specific skills
python3 evals/run_all.py --skills text-to-speech agents -v
# Custom model (see `cursor-agent --list-models`)
python3 evals/run_all.py --model gpt-5.4-high -vResults are saved to evals/results/<timestamp>/ with a report.md summary and results.json for programmatic access.
Functional evals use an isolated cursor-agent workspace per test case (under that results tree); they do not modify skill sources under each skill’s directory.
Requires the Cursor Agent CLI (cursor-agent on your PATH; override binary with CURSOR_AGENT) and Cursor authentication (cursor-agent login or CURSOR_API_KEY).
MIT