MCP Maigret

What BurtTheCoder/mcp-maigret MCP server does

BurtTheCoder/mcp-maigret MCP server brings the Maigret OSINT tool into MCP-compatible applications. It searches public websites and social networks for a supplied username, then produces a report in a selected format. It can also inspect a URL, extract information from it, and search for associated usernames.

The server is intended for legitimate research using publicly available information. Searches should respect privacy requirements, applicable data protection laws, and the terms of the services being queried. Automated requests may also be rate-limited or blocked by some sites.

How it works

The MCP interface exposes two primary operations. search_username accepts a username and optional search controls. By default, it searches the top 500 sites, but use_all_sites can expand the search to all available sites. A tags array can narrow the sites searched, such as sites tagged for photography, dating, or a region. Usernames are limited to letters and numbers, underscores, hyphens, and periods, with a maximum length of 100 characters.

parse_url accepts an HTTP or HTTPS URL and can generate a report while looking for related usernames. Both operations default to PDF output, with TXT, HTML, JSON, CSV, and XMind also supported. The listed MCP fields include username, format, use_all_sites, tags, url, and the corresponding operation names such as username, format, use_all_sites, tags, parse_url, and url.

Setup and configuration

BurtTheCoder/mcp-maigret MCP server requires Node.js 18 or newer, Docker, and a writable reports directory. Docker Desktop is required on Windows and macOS; Linux users need a working Docker installation and permission to run Docker commands.

The documented npm setup installs the package globally with npm install -g mcp-maigret. Configure the MCP client to run the mcp-maigret command and set MAIGRET_REPORTS_DIR to an existing directory where reports can be written. The README specifically documents Claude Desktop configuration and requires restarting the client after configuration changes. A source-based option clones the repository, runs npm install, builds the project, and starts its generated build/index.js file with Node.js.

Tools and capabilities

• Search usernames across hundreds of public sites.
• Expand searches beyond the default top-500 site set.
• Filter searched sites with validated tags.
• Parse profile or other HTTP/HTTPS URLs for associated usernames.
• Write reports as TXT, HTML, PDF, JSON, CSV, or XMind files.
• Execute Maigret using Docker for the underlying search process.

Limitations and notes

The reports directory must already exist and be writable, and the MAIGRET_REPORTS_DIR setting must be present. Docker must be installed and running. Invalid usernames, URLs, or tags are rejected according to the server’s input rules, including checks against shell metacharacters in URLs.

The implementation uses argument arrays and execFile() rather than shell-interpolated command strings, and Docker commands are run without shell interpretation. These safeguards reduce command-injection risk but do not change the need to use the server only for authorized, lawful research. Site availability and results can vary because external services may restrict automated access.