honeybadger

The Honeybadger MCP Server exposes Honeybadger's API through the MCP (Model Context Protocol) so you can interact with Honeybadger data via standard MCP clients. It runs in read-only mode by default, enabling safe access to projects and faults without allowing destructive changes unless you explicitly disable read-only. The server accepts your Honeybadger API token as an environment variable and forwards requests to Honeybadger's API, returning structured results that MCP-enabled tools can parse. Tools available through this server include listing projects, fetching project details, and managing projects (when read-only is disabled). Typical workflows involve configuring an MCP client (Cursor, Windsurf, Claude Code, VS Code, or Zed) to point at this server, then issuing commands such as list_projects, get_project, create_project, and update_project depending on your permission level.

Prerequisites:

• Docker installed on your machine (requires access to ghcr.io/honeybadger-io/honeybadger-mcp-server or building locally).
• Honeybadger API token with appropriate permissions.

1. Pull the official Docker image (recommended):

docker pull ghcr.io/honeybadger-io/honeybadger-mcp-server:latest

2. Configure your MCP client using the recommended docker-based server:

# Example mcp config (Cursor, Windsurf, Claude Desktop, VS Code, or Zed)
{
  "mcpServers": {
    "honeybadger": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "HONEYBADGER_PERSONAL_AUTH_TOKEN",
        "ghcr.io/honeybadger-io/honeybadger-mcp-server"
      ],
      "env": {
        "HONEYBADGER_PERSONAL_AUTH_TOKEN": "your personal auth token"
      }
    }
  }
}

3. If you prefer to run locally without pulling the image every time, you can run the image directly and supply your token:

docker run -i --rm -e HONEYBADGER_PERSONAL_AUTH_TOKEN=your_token ghcr.io/honeybadger-io/honeybadger-mcp-server

4. For building from source or alternative deployment methods, refer to the README examples for building from source and using the stdio mode as the command:

./honeybadger-mcp-server stdio --auth-token your_token --log-level debug --api-url https://app.honeybadger.io

Environment variables to be aware of:

• HONEYBADGER_PERSONAL_AUTH_TOKEN: required token for Honeybadger API access.
• HONEYBADGER_READ_ONLY: when set to false, write operations (create/update/delete) become available; defaults to true for safety.
• HONEYBADGER_API_URL: override Honeybadger API base URL if needed (default: https://app.honeybadger.io).
• LOG_LEVEL: controls logging verbosity (debug, info, warn, error).

Tips:

• The server is designed to be read-only by default to prevent accidental modifications. Explicitly set HONEYBADGER_READ_ONLY=false to enable mutating operations like creating or updating projects.
• If you switch to a local build, replace the image reference in your MCP client configuration with your local image tag or name.
• When using VS Code, Cursor, Windsurf, Claude Code, or Zed, ensure the token is provided securely (do not hard-code tokens in shared configs).