Evals Mcp Server

Overview

Evals Mcp Server lets you author verifiable evaluation records through a structured draft, review, revise, and submit workflow with enforced graders. By implementing the Model Context Protocol, it turns evaluation management into a set of callable tools that AI assistants can use directly, eliminating manual copy-paste between applications.

The server runs locally via the npm package @cyanheads/evals-mcp-server, communicating over stdio with no authentication required. Version v0.1.3 is available now under the Apache-2.0 license, with full source code at cyanheads/evals-mcp-server on GitHub.

Evals Mcp Server belongs in the AI & Agents category on mcp.site and works with MCP clients including Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, and Zed. If you maintain this server, claim your listing to earn the Verified badge and ensure the information stays accurate.

Installation

// claude_desktop_config.json
{
  "mcpServers": {
    "evals_mcp_server": {
      "command": "npx",
      "args": [
        "-y",
        "@cyanheads/evals-mcp-server"
      ]
    }
  }
}