Use the MCP Server in your IDE

Drive Future AGI from Claude, Cursor, or VS Code instead of the dashboard

What the MCP Server is

The Future AGI MCP Server lets you interact with the entire Future AGI platform through natural language, directly from your AI coding environment. Instead of switching between the dashboard and your editor, you can run evaluations, upload datasets, generate synthetic data, and apply protection rules just by describing what you want in tools like Claude, Cursor, or VS Code. It’s built on the Model Context Protocol, a standard that connects AI models to external tools and services.

How it works

Install

Add the MCP server to your IDE using the config below.

Authorize

OAuth login opens in your browser automatically. No API keys needed.

Start using

Ask your AI assistant about your evaluations, traces, datasets, and more.

Connect Your IDE

All you need is this URL. Authentication happens automatically via OAuth 2.0:

https://api.futureagi.com/mcp

Or add manually:

~/.cursor/mcp.json

{
  "mcpServers": {
    "futureagi": {
      "url": "https://api.futureagi.com/mcp"
    }
  }
}

Or add manually:

Run in terminal

claude mcp add futureagi --transport http https://api.futureagi.com/mcp

Or add manually:

.vscode/settings.json

{
  "mcp.servers": {
    "futureagi": {
      "type": "http",
      "url": "https://api.futureagi.com/mcp"
    }
  }
}

Add to your config file:

claude_desktop_config.json

{
  "mcpServers": {
    "futureagi": {
      "url": "https://api.futureagi.com/mcp"
    }
  }
}

Add to your config file:

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "futureagi": {
      "serverUrl": "https://api.futureagi.com/mcp"
    }
  }
}

What You Can Do

With Future AGI’s MCP Server, you can use natural language to:

Run automatic evaluations: evaluate batch and single inputs on various evaluation metrics, both on local datapoints and large datasets
Build and observe your agents: add observability, and evaluations while you build and deploy agents into production
Manage datasets: upload, evaluate, download datasets and find insights
Add protection rules: apply toxicity detection, prompt injection protection, and other guardrails automatically
Generate synthetic data: describe your dataset and objective to generate synthetic data

Check out our blog post on futureagi-mcp-server for detailed use cases.

Next Steps

Run your first evaluation using natural language through the MCP server
Explore the Observe quickstart to add tracing to your project
Learn about Protect to set up real-time guardrails for your AI application

Was this page helpful?

Questions & Discussion