Resemble AI MCP Server

MCP serverDev tools

This app gives your AI access to Resemble AI's media tools. Once it is added, your AI can check media for deepfakes, analyze media content, apply watermarks, and look up Resemble AI's documentation. It is built for development work involving media.

Use Resemble AI MCP Server in Claude, ChatGPT or Ahel Desktop

Free. Sign in, add Resemble AI MCP Server and connect your AI. About a minute.

Also: Claude Code · Cursor · Codex

Then ask your AI: use the detect deepfake tool from Resemble AI MCP Server

Details

Available today. Use it from your connected AI after setup.

After adding the app, ask your AI to run a deepfake check on a piece of media or to search the Resemble AI documentation for how a feature works.

Resemble AI MCP ServerStart free

What your AI can do with it

  • Detect deepfakes in media
  • Analyze media content for insights
  • Apply watermarks to media
  • Look up Resemble AI documentation

Install Resemble AI MCP Server

The server’s own address, for the clients that take one directly. Or connect Ahel once and every client you use reads it from one address, with the account kept on Ahel rather than in each client’s config.

  • Claude Code

    claude mcp add --transport http --scope user resemble-ai-mcp-server 'https://mcp.resemble.ai/mcp'

    Run it once in your project, then open /mcp to approve any sign-in the server asks for.

  • Claude Desktop

    https://mcp.resemble.ai/mcp

    Add a custom connector in Settings, paste this address, and approve the sign-in.

  • Cursor

    cursor://anysphere.cursor-deeplink/mcp/install?name=resemble-ai-mcp-server&config=eyJ1cmwiOiJodHRwczovL21jcC5yZXNlbWJsZS5haS9tY3AifQ==

    Open the link and Cursor adds the server at that address.

  • ChatGPT

    https://mcp.resemble.ai/mcp

    In Settings, enable Developer mode, create an MCP app, and paste this address. Your plan and workspace must allow custom apps.

  • Codex

    codex mcp add resemble-ai-mcp-server --url 'https://mcp.resemble.ai/mcp'

    Run it once, then sign in with codex mcp login resemble-ai-mcp-server if the server asks for an account.

From the project's README

As published by resemble-ai/resemble-mcp in README.md.

Official MCP server exposing Resemble AI documentation as resources and tools. Connect to Cursor, Claude Code, or any MCP-compatible client to enable your coding assistant to access Resemble AI docs during development.

Overview

This is the official Resemble AI MCP server. It implements the Model Context Protocol (stdio transport) to serve documentation from docs/pages/ as queryable resources. The server aggregates docs by topic and provides full-text search across all MDX files.

What it provides:

  • Topic-based documentation aggregation (13 topics: voice-cloning, text-to-speech, agents, etc.)
  • Full-text search across all MDX documentation pages
  • OpenAPI 3.0 spec with exact request/response schemas for code generation
  • Direct page and endpoint access

Once connected, your coding assistant can query Resemble AI documentation programmatically. You can vibe code complete applications—your assistant has instant access to guides, tutorials, and exact API schemas for accurate code generation.

Quick Start

1. Clone & Install

git clone https://github.com/resemble-ai/resemble-mcp.git
cd resemble-mcp

# Create virtual environment
python3 -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

2. Configure MCP Client

Cursor: Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "resemble-ai-docs": {
      "command": "/path/to/resemble-mcp/venv/bin/python",
      "args": ["/path/to/resemble-mcp/server.py"]
    }
  }
}

Claude Code / Claude Desktop: Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or equivalent location for your OS:

{
  "mcpServers": {
    "resemble-ai-docs": {
      "command": "/path/to/resemble-mcp/venv/bin/python",
      "args": ["/path/to/resemble-mcp/server.py"]
    }
  }
}

Note: Use absolute paths. Replace /path/to/resemble-mcp with your actual path.

3. Restart Client

Restart your MCP client to load the server. Tools will be available after restart.

Available Tools

resemble_docs_lookup

Returns aggregated documentation for a topic. Accepts topic ID and returns all related pages concatenated.

Parameters:

  • topic (string, required): One of: voice-cloning, text-to-speech, speech-to-speech, speech-to-text, getting-started, voice-design, streaming, detect, agents, sdks, projects-clips, audio-tools, ssml

Returns: Markdown content with all pages for the topic.

resemble_search

Full-text search across all documentation pages. Returns matching pages with snippets.

Parameters:

  • query (string, required): Search query
  • limit (integer, optional): Max results (default: 5)

Returns: List of matching pages with titles, paths, and content snippets.

resemble_get_page

Retrieves a specific documentation page by path.

Parameters:

  • path (string, required): Page path relative to docs/pages/ (e.g., getting-started/quickstart)

Returns: Full page content in Markdown.

resemble_list_topics

Lists all available topic IDs with descriptions.

Returns: Markdown list of topics.

resemble_api_endpoint

Get detailed OpenAPI specification for a specific API endpoint. Returns exact request/response schemas, parameters, and types.

Parameters:

  • method (string, required): HTTP method (GET, POST, PUT, PATCH, DELETE)
  • path (string, required): API endpoint path (e.g., /synthesize, /voices)

Returns: Endpoint specification with schemas, parameters, and response formats.

resemble_api_search

Search the OpenAPI spec for endpoints matching a query.

Parameters:

  • query (string, required): Search query (e.g., voice, streaming, agent)

Returns: List of matching endpoints with methods, paths, and summaries.

Usage

Once configured, your coding assistant (Cursor, Claude Code, etc.) can call these tools directly. Example use cases:

Building a voice cloning app:

{
  "name": "resemble_docs_lookup",
  "arguments": { "topic": "voice-cloning" }
}

Searching for streaming implementation:

{
  "name": "resemble_search",
  "arguments": { "query": "websocket streaming", "limit": 5 }
}

Getting exact API schema for implementation:

{
  "name": "resemble_api_endpoint",
  "arguments": { "method": "POST", "path": "/synthesize" }
}

Finding all voice-related API endpoints:

{
  "name": "resemble_api_search",
  "arguments": { "query": "voice" }
}

Your assistant uses these tools automatically when you ask it to build features. For example, "build me a TTS streaming service" triggers resemble_docs_lookup for the text-to-speech topic, and may use resemble_api_endpoint to get exact request schemas for accurate implementation.

Project Structure

resemble-mcp/
├── server.py               # MCP server implementation
├── docs/
│   └── pages/              # MDX documentation files
│       ├── getting-started/    # Auth, quickstart, basics
│       ├── voice-generation/   # TTS, STT, S2S
│       ├── voice-creation/     # Voice cloning, recordings
│       ├── agents/             # AI agents
│       ├── detect/             # Deepfake detection
│       └── guides/             # Step-by-step tutorials
└── openapi/
    └── openapi.yml         # OpenAPI 3.0 specification

Supported MCP Clients

Works with any client implementing the Model Context Protocol:

Development

The server uses stdio transport (reads from stdin, writes to stdout) as per MCP specification.

# Run server directly (for testing/debugging)
python server.py

Architecture:

  • Server implementation: server.py
  • Documentation source: docs/pages/ (MDX files)
  • API specification: openapi/openapi.yml (OpenAPI 3.0)
  • Topic definitions: Hardcoded in TOPICS dict
  • Search: Full-text search with scoring across docs and OpenAPI endpoints

Troubleshooting

Server not loading:

  • Verify absolute paths in MCP config
  • Ensure virtual environment is activated
  • Check mcp package is installed: pip show mcp

Tools not appearing:

  • Restart MCP client
  • Check client logs for initialization errors
  • Verify server starts without errors: python server.py

Search returns no results:

  • Use resemble_docs_lookup with topic ID for guaranteed results
  • Try broader search terms
  • Check topic list: resemble_list_topics

License

MIT License - see LICENSE for details.

Links

Tools it offers (18)

What this server listed when Ahel dialed its public endpoint in Oct 2026, with no key and no account of yours. The names are the server’s own.

  • detect_deepfake
  • get_detection
  • detect_ai_text
  • get_text_detection
  • analyze_media
  • ask_about_detection
  • detect_watermark
  • apply_watermark
  • trace_audio_source
  • list_detect_agents
  • run_detect_agent_investigation
  • get_detect_agent_run
  • agent_detection_list_sites
  • agent_detection_create_site
  • agent_detection_update_site
  • agent_detection_get_analytics
  • agent_detection_list_sessions
  • agent_detection_get_session

Signals

GitHub stars
1
Last commit
Sep 2026
Advanced
Delivery
resemble-mcp MCP server → your Ahel connector (mcp.ahel.ai) → your AI.
Item type
mcp-server
Key
io-github-resemble-ai-resemble-mcp
Source
github.com/resemble-ai/resemble-mcp
Hosted endpoint
https://mcp.resemble.ai/mcp