Overview
What is drand-mcp-server?
drand-mcp-server is a Model Context Protocol (MCP) server that lets AI applications fetch verifiable randomness from the drand network. It provides tools to get the latest beacon, a beacon by round, or a beacon by timestamp, and use that randomness as an input seed in model-driven flows.
How to use drand-mcp-server?
Run the server via npx drand-mcp-server or build from source (npm install && npm run build) and then run node ./dist/index.mjs. Configure it as an MCP server in VS Code (.vscode/mcp.json) or Claude Desktop (claude_desktop_config.json) by setting the command and args as shown in the README.
Key features of drand-mcp-server
- Fetch the latest random beacon from drand Quicknet.
- Retrieve a beacon by a specific round number.
- Get the beacon emitted at or just before a given timestamp.
- All randomness is cryptographically verifiable.
- Built for integration with AI tools via the MCP protocol.
Use cases of drand-mcp-server
- Repeatable, random sampling of input data.
- Verifiable interaction with other MCP servers (e.g., paying out rewards based on a prompt).
- Verifying the output of another random process using historical drand beacons.
FAQ from drand-mcp-server
What runtime does drand-mcp-server require?
Node.js version 21 or later (the fetch API is required).
How do I obtain a randomness beacon for a specific time?
Use the get-randomness-by-time tool with a time parameter in milliseconds; it returns the beacon emitted at or just before that time.
Does drand-mcp-server store any data locally?
No. The server fetches beacons directly from the drand network and does not persist data locally.
Which transport protocol does drand-mcp-server use?
The README describes command‑line execution (stdio) for VS Code and Claude Desktop; no other transport (e.g., HTTP or WebSocket) is mentioned.
Is authentication required to use drand-mcp-server?
No authentication or API keys are mentioned in the README. The server connects to the public drand Quicknet.
