ElevenLabs MCP Server
A text-to-speech MCP server that turns text and multi-voice scripts into audio through ElevenLabs and keeps a job history.
- 119 ★ on the repository
- Last commit
Extended access
Why this level
- Voiceover text is sent to an external paid service under your key.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/elevenlabs-mcp-serverDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Voiceover text is sent to an external paid service under your key.
This is third-party code. Review the repository files before installing.
What it does
The server connects an agent to the ElevenLabs speech synthesis API. The `generate_audio_simple` tool voices plain text with the default voice, while `generate_audio_script` builds audio from a script with several voices and roles. Job history is stored in SQLite so voiceovers can be listed, downloaded or deleted with `get_voiceover_history`, `get_audio_file` and `delete_job`. `list_voices` returns available voices. Voice and model settings are set through environment variables: voice ID, model, stability, similarity boost, style and output directory. The repository also contains a sample SvelteKit web client.
Who it is for. Creators of videos, podcasts and learning material who voice scripts through an agent and want a history of voiceovers.
Good fit when
- A video script with two hosts must become a single audio track.
- An agent voices short lines for a voice interface prototype.
- Earlier voiceovers need to be found by job ID and downloaded again.
- You need to switch voice and model through environment variables with no code changes.
Not a fit when
- You have no ElevenLabs key or cannot send text to an external service.
- You need local offline speech.
- You need real-time speech: the server produces audio files rather than streamed sound.
Example request
Voice this script with two voices, save the file and show the job ID in the history.MCP
- Transport
- stdio
- Authentication
- API key
Environment variables
`ELEVENLABS_API_KEY`required, secret- ElevenLabs API key.
`ELEVENLABS_VOICE_ID`- Default voice ID.
`ELEVENLABS_MODEL_ID`- Synthesis model, `eleven_flash_v2` in the example.
`ELEVENLABS_OUTPUT_DIR`- Directory for generated audio files.
Limitations
You need your own `ELEVENLABS_API_KEY`, and synthesis is billed by ElevenLabs. Voiceover text goes to an external service. Generated files and history stay in a local directory and SQLite, which matters for private material. The web client in the repository is a sample installed separately with `pnpm`. Last commit October 2026. The Smithery install configures the Claude Desktop client automatically, while for other clients the `mcpServers` block is added by hand.
How to disable. Remove the `elevenlabs` entry from MCP settings, revoke the key in the ElevenLabs account and clear the output directory and history database.
FAQ
How do I connect without Smithery?
Add a `uvx` command with the `elevenlabs-mcp-server` argument and an `env` block with the key and voice settings to MCP settings. No separate install is needed.
Which voice settings exist?
`ELEVENLABS_VOICE_ID`, `ELEVENLABS_MODEL_ID`, `ELEVENLABS_STABILITY`, `ELEVENLABS_SIMILARITY_BOOST`, `ELEVENLABS_STYLE` and `ELEVENLABS_OUTPUT_DIR`.
How do I voice a script with several roles?
Call `generate_audio_script` with a structured script of voices and actors. The result is stored in the job history.
Where is history stored?
In a SQLite database inside the server. Read it with `get_voiceover_history`, while the `voiceover://history/{job_id}` resource returns a file by job ID.
What licence does the project use?
MIT. The repository can be used and modified under that licence, while speech synthesis itself is billed by ElevenLabs.
Discussion
No questions or reviews yet.
Related
MCP serversA local MCP server for voice work on macOS: voice cloning, speech synthesis and audiobooks from PDF on an agent's command
Local speech synthesis for agents: a command and a skill voice text, clone a voice from a recording and render scores to audio, with no cloud
An MCP server for Sber's SaluteSpeech voice API: records speech from the microphone, recognizes it and speaks the agent's reply
A local sound studio with an MCP server: an agent makes effects, music, songs and speech on open-weight models
Official MiniMax MCP server: text to speech, voice cloning and design, image and video generation
MCP server for Sber SaluteSpeech: transcribe Russian speech to text and voice text with Sber voices
README in short
The recommended connection goes through `uvx`: no separate install is needed, and the key and voice parameters are passed in an `env` block in MCP settings.