Skip to main content
Glama

elevenlabs-mcp

A full-featured Model Context Protocol (MCP) server for the ElevenLabs API — bring text‑to‑speech, speech‑to‑text, voice cloning, voice design, sound effects, music, audio isolation, dubbing, and account tools directly into Claude Code (or any MCP client).

Built in Python with FastMCP and the official elevenlabs SDK (2.x), managed with uv.

Audio‑producing tools save their output to a folder on disk (your Desktop by default) and return the file path. Data tools return JSON. Your API key is read from the environment and is never written to the repo.


Features

28 tools across the ElevenLabs surface:

Category

Tools

🔊 Speech generation

text_to_speech, text_to_dialogue (multi‑speaker)

🎚️ Voice transformation

speech_to_speech (voice changer)

🎵 Audio generation

sound_effects, compose_music

🧹 Audio cleanup

audio_isolation (remove background noise)

📝 Transcription

speech_to_text (Scribe), forced_alignment

🧬 Voice design & cloning

design_voice, create_voice_from_preview, clone_voice (instant)

🗂️ Voice management

list_voices, get_voice, search_voice_library, get_default_voice_settings, edit_voice_settings, delete_voice

🧠 Models & account

list_models, get_user_info, check_subscription, get_usage_stats

🕓 History

list_history, get_history_item, download_history_audio, delete_history_item

🌍 Dubbing

dub_audio, get_dubbing_status, download_dubbed_audio


Related MCP server: ElevenLabs MCP Server

Requirements


Installation

git clone https://github.com/neomorrison/elevenlabs-mcp.git
cd elevenlabs-mcp
uv sync

uv sync creates a virtual environment and installs the server plus its dependencies.

Add to Claude Code

Register the server with the Claude Code CLI. Replace the path with wherever you cloned the repo, and paste your real key:

Windows (PowerShell):

claude mcp add elevenlabs -s user -e ELEVENLABS_API_KEY=sk_your_key_here -- uv run --directory C:\path\to\elevenlabs-mcp elevenlabs-mcp

macOS / Linux:

claude mcp add elevenlabs -s user -e ELEVENLABS_API_KEY=sk_your_key_here -- uv run --directory /path/to/elevenlabs-mcp elevenlabs-mcp

Then verify:

claude mcp list

You should see elevenlabs listed as Connected. Restart Claude Code (or start a new session) and the tools become available.

-s user installs the server for your user across all projects. Use -s local (the default) to scope it to the current project, or -s project to share it via a checked‑in .mcp.json.

Other MCP clients (Claude Desktop, etc.)

Add this to your client's MCP config (e.g. claude_desktop_config.json):

{
  "mcpServers": {
    "elevenlabs": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/elevenlabs-mcp", "elevenlabs-mcp"],
      "env": { "ELEVENLABS_API_KEY": "sk_your_key_here" }
    }
  }
}

Configuration

Configuration is via environment variables (passed with -e in claude mcp add, or in a local .env for development — see .env.example):

Variable

Required

Default

Description

ELEVENLABS_API_KEY

✅

—

Your ElevenLabs API key.

ELEVENLABS_MCP_OUTPUT_DIR

~/Desktop

Where generated audio/media files are written. (ELEVENLABS_MCP_BASE_PATH is accepted as an alias.)

ELEVENLABS_DEFAULT_VOICE_ID

JBFqnCBsd6RMkjVDRZzb (George)

Voice used when a tool call omits voice_id.

ELEVENLABS_DEFAULT_MODEL_ID

eleven_multilingual_v2

Default TTS model.


Usage

Once installed, just ask Claude in natural language. Examples:

  • “Use ElevenLabs to read this paragraph in George's voice and save the mp3.”

  • “List my ElevenLabs voices.”

  • “Generate a 10‑second sound effect of rain on a tin roof.”

  • “Transcribe C:\\recordings\\interview.wav with speaker diarization.”

  • “Design a voice: a calm, wise elderly narrator, then save the second preview.”

  • “Clone a voice named ‘Studio’ from these three sample files.”

  • “How many characters do I have left this month?”

  • “Dub intro.mp4 into Spanish.”

Generated files land in your output directory (Desktop by default); each tool returns the absolute path.


Development

uv sync
# run the server directly over stdio (Ctrl+C to stop)
uv run elevenlabs-mcp

Inspect the tools with the MCP Inspector:

uv run mcp dev src/elevenlabs_mcp/server.py

Project layout:

src/elevenlabs_mcp/
  __init__.py
  server.py        # FastMCP server + all tools

Notes & limitations

  • Most audio tools make paid ElevenLabs API calls — usage counts against your plan.

  • Dubbing is asynchronous: call dub_audio, poll get_dubbing_status until status is dubbed, then download_dubbed_audio.

  • The server never prints to stdout (that channel belongs to the MCP transport); logs go to stderr.

License

MIT © neomorrison

Available Tools

28 tools
audio_isolationA

Isolate speech from an audio file by removing background noise/music.

Args: audio_file_path: path to the source audio file. output_filename: optional output file name.

Returns the absolute path of the cleaned audio file.

Inputs: audio_file_path (required), output_filename. Returns: result

check_subscriptionA

Check subscription tier and character usage/remaining for the account.

Inputs: no named parameters. Returns: result

clone_voiceA

Instant voice clone: create a new voice from one or more audio samples.

Args: name: name for the cloned voice. audio_file_paths: paths to clean sample recordings of the target voice. description: optional description. labels: optional dict of metadata labels. remove_background_noise: clean the samples before cloning.

Returns the new voice's id as JSON.

Inputs: name (required), audio_file_paths (required), labels, description, +1 more. Returns: result

compose_musicA

Compose music from a text prompt (ElevenLabs Music).

Args: prompt: description of the music, e.g. "upbeat lofi hip hop, mellow piano". music_length_ms: desired length in milliseconds (e.g. 30000 for 30s). model_id: music model id (API default when omitted). output_format: audio output format. output_filename: optional output file name.

Returns the absolute path of the saved audio file.

Inputs: prompt (required), model_id, output_format, music_length_ms, +1 more. Returns: result

create_voice_from_previewA

Save a previously designed voice preview as a permanent voice.

Args: voice_name: name for the new voice. voice_description: description of the voice. generated_voice_id: id returned by design_voice.

Returns the new voice's id and name as JSON.

Inputs: voice_name (required), voice_description (required), generated_voice_id (required). Returns: result

delete_history_itemB

Delete a history item.

Args: history_item_id: the history item id to delete.

Inputs: history_item_id (required). Returns: result

delete_voiceC

Delete a voice from your account.

Args: voice_id: the voice id to delete.

Inputs: voice_id (required). Returns: result

design_voiceA

Design candidate voices from a text description (Voice Design).

Generates preview voices and saves each preview's audio to a file. To keep a preview permanently, call create_voice_from_preview with its generated_voice_id.

Args: voice_description: description of the desired voice, e.g. "an old British narrator with a warm, gravelly tone". text: sample text to speak in the previews (>=100 chars recommended). model_id: voice-design model id (API default when omitted). auto_generate_text: let ElevenLabs generate the preview text.

Returns JSON listing each preview's generated_voice_id and saved audio path.

Inputs: voice_description (required), text, model_id, auto_generate_text. Returns: result

download_dubbed_audioA

Download the dubbed audio for a completed dubbing job.

Args: dubbing_id: the dubbing job id. language_code: the target language code to download. output_filename: optional output file name.

Returns the absolute path of the saved audio file.

Inputs: dubbing_id (required), language_code (required), output_filename. Returns: result

download_history_audioB

Download the audio for a history item and save it to a file.

Args: history_item_id: the history item id. output_filename: optional output file name.

Returns the absolute path of the saved audio file.

Inputs: history_item_id (required), output_filename. Returns: result

dub_audioA

Start a dubbing job to translate audio/video into another language.

Provide either a local file or a source URL (e.g. YouTube). Dubbing runs asynchronously — poll with get_dubbing_status and then download_dubbed_audio.

Args: target_lang: target language code (e.g. "es", "fr", "de"). audio_file_path: path to a local audio/video file. source_url: URL of the source media (alternative to a file). source_lang: source language code (auto-detected if omitted). name: optional project name. num_speakers: expected number of speakers (0 = auto). watermark: apply ElevenLabs watermark.

Returns the dubbing_id as JSON.

Inputs: target_lang (required), name, watermark, source_url, +3 more. Returns: result

edit_voice_settingsA

Update the settings for a voice (only provided fields are changed).

Args: voice_id: the voice id. stability / similarity_boost / style / speed: floats (0..1; speed ~0.7..1.2). use_speaker_boost: bool.

Returns the updated settings as JSON.

Inputs: voice_id (required), speed, style, stability, +2 more. Returns: result

forced_alignmentA

Align a known transcript to audio, returning per-word/character timings.

Args: audio_file_path: path to the audio file. text: the exact transcript of the audio.

Returns alignment data as JSON.

Inputs: text (required), audio_file_path (required). Returns: result

get_default_voice_settingsA

Return the default voice settings used by ElevenLabs.

Inputs: no named parameters. Returns: result

get_dubbing_statusC

Get the status and metadata of a dubbing job.

Args: dubbing_id: the dubbing job id.

Inputs: dubbing_id (required). Returns: result

get_history_itemC

Get metadata for a single history item.

Args: history_item_id: the history item id.

Inputs: history_item_id (required). Returns: result

get_usage_statsA

Get character/credit usage statistics over a time window.

Args: start_unix: window start (unix seconds). Defaults to 30 days ago. end_unix: window end (unix seconds). Defaults to now. metric: one of "credits", "tts_characters", "minutes_used", "request_count". breakdown_type: e.g. "none", "voice", "model", "api_keys". aggregation_interval: "hour", "day", "week", "month", or "cumulative".

Returns usage data as JSON.

Inputs: metric, end_unix, start_unix, breakdown_type, +1 more. Returns: result

get_user_infoA

Get information about the authenticated ElevenLabs user.

Inputs: no named parameters. Returns: result

get_voiceB

Get full details for a single voice, including its settings.

Args: voice_id: the voice id.

Returns the voice object as JSON.

Inputs: voice_id (required). Returns: result

list_historyB

List recent generation history items.

Args: page_size: number of items to return. voice_id: filter to a specific voice. search: free-text search over history.

Returns a JSON summary of history items.

Inputs: search, voice_id, page_size. Returns: result

list_modelsA

List the available ElevenLabs models and their capabilities.

Inputs: no named parameters. Returns: result

list_voicesA

List the voices available in your account.

Args: search: optional case-insensitive substring to filter by name/labels.

Returns a JSON summary of each voice (id, name, category, labels).

Inputs: search. Returns: result

search_voice_libraryA

Search the shared ElevenLabs voice library (community voices).

Args: search: free-text search. category: e.g. "professional", "high_quality", "famous". gender: e.g. "male", "female". language: language filter. page_size: max results to return.

Returns a JSON summary of matching shared voices.

Inputs: gender, search, category, language, +1 more. Returns: result

sound_effectsA

Generate a sound effect from a text prompt.

Args: text: description of the sound effect, e.g. "distant thunder rolling". duration_seconds: desired length (0.5..30). If omitted the model chooses. prompt_influence: 0..1 — how strictly to follow the prompt. loop: make the effect loopable. output_format: audio output format. output_filename: optional output file name.

Returns the absolute path of the saved audio file.

Inputs: text (required), loop, output_format, output_filename, +2 more. Returns: result

speech_to_speechA

Voice changer: transform speech in an audio file to a target voice.

Args: audio_file_path: path to the source audio file. voice_id: target voice id (defaults to the configured default voice). model_id: STS model id (default "eleven_multilingual_sts_v2"). output_format: audio output format. remove_background_noise: strip background noise before conversion. output_filename: optional output file name.

Returns the absolute path of the converted audio file.

Inputs: audio_file_path (required), model_id, voice_id, output_format, +2 more. Returns: result

speech_to_textA

Transcribe an audio file to text (ElevenLabs Scribe).

Args: audio_file_path: path to the audio file to transcribe. model_id: STT model id (default "scribe_v1"). language_code: optional ISO language hint. diarize: annotate which speaker said what. tag_audio_events: tag non-speech events like (laughter). num_speakers: hint for the expected number of speakers.

Returns the transcription result as JSON (text, language, words with timings).

Inputs: audio_file_path (required), diarize, model_id, num_speakers, +2 more. Returns: result

text_to_dialogueA

Generate multi-speaker dialogue audio from a list of turns.

Args: dialogue: list of turns, each {"voice_id": "...", "text": "..."}. model_id: dialogue model id (defaults to "eleven_v3" style dialogue model when omitted, letting the API choose). output_format: audio output format. output_filename: optional output file name.

Returns the absolute path of the saved audio file.

Inputs: dialogue (required), model_id, output_format, output_filename. Returns: result

text_to_speechA

Convert text to speech and save the audio to a file.

Args: text: The text to synthesize. voice_id: ElevenLabs voice id (defaults to the configured default voice). model_id: TTS model id (default "eleven_multilingual_v2"). Other options include "eleven_flash_v2_5" (fast/cheap) and "eleven_turbo_v2_5". output_format: e.g. "mp3_44100_128", "mp3_44100_192", "pcm_44100", "wav_44100". language_code: ISO code to enforce a language (supported by some models). stability / similarity_boost / style / speed: voice setting overrides (0..1; speed ~0.7..1.2). use_speaker_boost: bool. seed: deterministic generation seed. output_filename: optional file name (extension inferred if omitted).

Returns the absolute path of the saved audio file.

Inputs: text (required), seed, speed, style, +8 more. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.7/5.0

    Scored across 28 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose: text generation, voice cloning, dubbing, transcription, sound effects, music composition, etc. No significant overlap; even similar-sounding tools like text_to_speech, text_to_dialogue, and speech_to_speech are well-differentiated by their descriptions.

    Naming Consistency5/5

    All tools use consistent snake_case naming with a verb_noun pattern (e.g., list_voices, delete_history_item, download_dubbed_audio). Minor deviations like forced_alignment are still in the same style, and there is no mixing of conventions like camelCase.

    Tool Count4/5

    28 tools is on the higher side but appropriate given the breadth of ElevenLabs features (TTS, voice management, dubbing, sound effects, music, etc.). Each major capability has its own set of tools, so the count feels comprehensive rather than bloated.

    Completeness4/5

    The tool surface covers most key workflows: voice CRUD, generation (TTS, STS, sound effects, music), dubbing (start, status, download), transcription, alignment, and usage stats. Minor gaps exist (e.g., no list_dubbing_jobs tool, no rename_voice), but agents can work around them.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      B
      quality
      F
      maintenance
      An official Model Context Protocol (MCP) server that enables AI clients to interact with ElevenLabs' Text to Speech and audio processing APIs, allowing for speech generation, voice cloning, audio transcription, and other audio-related tasks.
      27
      1,551
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      MCP server that brings ElevenLabs to Claude Code — text-to-speech, sound effects, music generation, voice cloning, speech-to-speech, transcription, and voice isolation. 8 tools for industry-leading AI audio.
      8
      MIT