Log inSubmit server
Elevenlabs MCP logo

Elevenlabs MCP

RunAPI MCP access to ElevenLabs audio generation, speech recognition, task polling, and pricing.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Elevenlabs MCP

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "elevenlabs-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@runapi.ai/elevenlabs-mcp"
      ],
      "env": {
        "RUNAPI_API_KEY": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

Automated check passed— started and listed 5 tools correctly (4d ago).

Overview

The elevenlabs-mcp MCP server exposes RunAPI's ElevenLabs models for audio isolation, speech-to-text, dialogue, sound effects, and text-to-speech. It creates asynchronous tasks, can wait for terminal results, and returns task IDs, statuses, and output URLs. Use it when an MCP client needs scoped ElevenLabs access rather than the full RunAPI model catalog. Pricing lookup does not require authentication, while task creation and polling require RunAPI credentials.

Use cases

  • Generate speech from text through ElevenLabs
  • Convert uploaded speech to text
  • Create dialogue or sound-effect tasks
  • Isolate audio from a media input
  • Check pricing before submitting a task

Key features

  • ElevenLabs audio task creation
  • Optional terminal-status waiting
  • Task status and payload retrieval
  • Output URL reporting
  • Current model and endpoint pricing lookup
  • RunAPI API-key and login-file authentication

Tools5

What Elevenlabs MCP lets your agent do. Expand a tool for its parameters.

Verified live~876 tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 3d ago

Recent health check succeeded.

7/7 checks healthy over the last 46d

What it can do

From the 5 tools this server returned when we ran it.

Handles credentials · 1Makes network requests · 1Read-only tools · 5
Network at startup3 connections
Files written at startupNone
Install scriptsNone
npm provenanceNot published
Package age118 days
Maintainers1

Observed 2026-10-20 in an isolated sandbox. A description of reach, not a security audit.

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 11d ago via OSV.dev · @runapi.ai/elevenlabs-mcp (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

npm downloads551Package downloads in the last 30 days.
Last commit9d agoMost recent push to the default branch.
Availability100%Our rolling endpoint + install checks that succeeded.
Install checkPassedOur sandbox started it and listed its tools.
Tools exposed5Callable tools this server registers over MCP.
Directory activity2 viewsConfig copies, upvotes, and views on AllMCPs.

How Elevenlabs MCP works

What elevenlabs-mcp MCP server does

The elevenlabs-mcp MCP server connects MCP-compatible assistants to the ElevenLabs model line through RunAPI. It focuses on five operation types: isolating audio, converting speech to text, generating dialogue from text, creating sound effects from text, and producing speech from text. The server also exposes task inspection and pricing lookup, so an agent can manage an operation from submission through result retrieval.

Five endpoints cover six listed model variants. Audio isolation uses audio-isolation; speech recognition uses speech-to-text; dialogue generation uses text-to-dialogue-v3; sound generation uses sound-effect-v2; and text-to-speech supports text-to-speech-multilingual-v2 and text-to-speech-turbo-v2.5. Availability may change between releases, so current pricing or the RunAPI model page should be treated as the current catalog reference.

How it works

The elevenlabs-mcp MCP server uses asynchronous RunAPI tasks. A creation tool submits the requested operation and can either return immediately or wait for a terminal status. Immediate submission returns a task ID for later inspection with get_task; waiting can return the task ID, current status, and output URLs in one interaction. get_task retrieves the latest status and payload for an existing task.

An assistant can call check_pricing before creating work. That tool accepts a model and endpoint lookup and works without signing in. The other operation tools and task polling require authentication. This makes the server suitable for workflows that compare current pricing before spending credits, submit media work, and check results later.

Setup and configuration

Install the elevenlabs-mcp MCP server as a standard stdio process with the npm package @runapi.ai/elevenlabs-mcp. Claude Code can register it with claude mcp add elevenlabs -s user -- npx -y @runapi.ai/elevenlabs-mcp; a project-scoped registration is also supported. Other MCP hosts can use npx as the command and -y @runapi.ai/elevenlabs-mcp as its arguments.

For authenticated operations, the server checks credentials in this order: RUNAPI_API_KEY, the RunAPI configuration file at ~/.config/runapi/config.json, then no key. The MCP login tool or runapi login can create the configuration file. Headless and CI environments can set RUNAPI_API_KEY before starting the host, or provide a configuration file containing an apiKey value. Do not commit actual keys to a repository.

Tools and capabilities

  • isolate_audio: submits an ElevenLabs audio-isolation task and can wait for completion.
  • speech_to_text: submits a speech-recognition task and can wait for completion.
  • text_to_dialogue: submits a text-to-dialogue task and can wait for completion.
  • text_to_sound: submits a text-to-sound task and can wait for completion.
  • text_to_speech: submits a text-to-speech task and can wait for completion.
  • get_task: retrieves the latest state and payload for an existing task.
  • check_pricing: looks up current pricing for a model and endpoint without authentication.

The creation tools return task metadata and output URLs when available. An agent can request wait: false when it needs to continue other work and use the returned ID for a later status check.

Limitations and notes

The server is scoped to ElevenLabs models available through RunAPI; it is not the full RunAPI catalog. Task creation and status retrieval need authentication, while pricing lookup is the unauthenticated exception. Results depend on asynchronous task state, so a submitted task may require a later get_task call.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Elevenlabs MCP

elevenlabs-mcp is an MCP server that connects AI assistants to ElevenLabs models through RunAPI. Its main tools create audio-isolation, speech-to-text, dialogue, sound-effect, and text-to-speech tasks, retrieve task status, and check current pricing. Authenticated operations return task IDs, statuses, and output URLs.

Install it with `npx -y @runapi.ai/elevenlabs-mcp` in an MCP host configuration. Claude Code can use `claude mcp add elevenlabs -s user -- npx -y @runapi.ai/elevenlabs-mcp`; Cursor, Windsurf, VS Code, Roo Code, Codex, and other MCP hosts can use the same stdio command.

Yes, elevenlabs-mcp works with Cursor through its standard stdio configuration. Set the command to `npx` and the arguments to `-y` and `@runapi.ai/elevenlabs-mcp`, then provide RunAPI authentication for task creation and polling.

Task creation and status polling require RunAPI authentication, while `check_pricing` works without a key. The server can read `RUNAPI_API_KEY` or credentials stored in `~/.config/runapi/config.json`, which can be created through the MCP `login` tool or `runapi login`.

Call `get_task` with the task ID returned by a creation tool. Submit the original operation with `wait: false` when you want an immediate response, then use `get_task` to retrieve the current status and latest payload.

Related MCP servers

All in Text-to-Speech All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewElevenlabs MCP AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/elevenlabs-mcp?style=directory)](https://allmcps.com/mcp/elevenlabs-mcp)
HTML Embed
<a href="https://allmcps.com/mcp/elevenlabs-mcp"><img src="https://allmcps.com/api/badge/elevenlabs-mcp?style=directory" alt="Elevenlabs MCP on AllMCPs" /></a>