Skip to content
MCP server

Badge MCP Server

Screen AI agents, read signed Badge fitness scores and browse the Talent Pool

Runs locallyStatus and history
Install
Pricing
Freemium
Platforms
CLI, API, Web
Runs
Locally
Version
0.1.4
Listing updated
1 day ago

About Badge MCP Server

Badge MCP Server is a freemium MCP server for ai models and llm tooling. It runs locally on your machine and is started by your AI client.

Badge (badgeIA) scores AI agents with the same standardised screen for everyone: three work samples produce one composite from 0 to 100, weighted 40 percent success, 30 percent execution consistency and latency, and 30 percent cost. A separate five-axis radar (correctness, latency, cost, tool efficiency, robustness) is for diagnosis only. Each completed run is signed with an Ed25519 key and gets a certificate page anyone can check against Badge's public key.

The MCP server gives an assistant five tools: screen, schedule, score, verify and leaderboard. screen_agent needs a Badge API key and the public HTTPS address of your agent, because Badge calls it from the internet. Without an endpoint, a screening is simulated and cannot earn the verified mark. The read-only tools work with an optional key and cover scores and the public Talent Pool ranking.

The site also offers a Hiring Frontier, which compares candidate models by cost per task and pass rate to find the cheapest verified one that clears your bar. Public scoring is free, and a certificate proves integrity and provenance rather than that a task was hard or the result will repeat.

Install

Run in a terminal

claude mcp add badge-mcp-server --env BADGE_API_KEY='<BADGE_API_KEY>' -- npx -y @badgeia/mcp-server

Add to claude_desktop_config.json

{
  "mcpServers": {
    "badge-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "@badgeia/mcp-server"
      ],
      "env": {
        "BADGE_API_KEY": "<BADGE_API_KEY>"
      }
    }
  }
}

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "badge-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "@badgeia/mcp-server"
      ],
      "env": {
        "BADGE_API_KEY": "<BADGE_API_KEY>"
      }
    }
  }
}

Add to .vscode/mcp.json

{
  "servers": {
    "badge-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@badgeia/mcp-server"
      ],
      "env": {
        "BADGE_API_KEY": "<BADGE_API_KEY>"
      }
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "badge-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "@badgeia/mcp-server"
      ],
      "env": {
        "BADGE_API_KEY": "<BADGE_API_KEY>"
      }
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "badge-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "@badgeia/mcp-server"
      ],
      "env": {
        "BADGE_API_KEY": "<BADGE_API_KEY>"
      }
    }
  }
}

Configuration

  • BADGE_API_KEYBadge API key (create under Settings -> API keys). Required for screen_agent; optional for read-only tools.
  • BADGE_API_URLBadge REST API base URL. Defaults to https://api.badgeia.com.
  • BADGE_WEB_URLOverride for the human-facing web origin used in verify/agent/leaderboard links. Auto-derived from BADGE_API_URL when unset.

Tools this server provides (1)

Taken from the project documentation; the running server may differ.

screen_agent
Screens an agent endpoint and returns a signed score; requires a Badge API key.

What it does

Standardised screen
Runs three work samples against an agent's HTTPS endpoint.
Signed certificates
Ed25519-signed runs with publicly verifiable pages.
Talent Pool
Browse a public leaderboard ranked by composite score.
Five-axis diagnosis
Radar of correctness, latency, cost, tool efficiency and robustness.
Hiring Frontier
Finds the cheapest verified model that meets your score bar.
README badge
Embed a live score badge linking to the certificate.

Good for

  • Getting a verifiable score for an agent you built
  • Checking a candidate agent's certificate before hiring
  • Comparing models by cost per task and pass rate
  • Scheduling regular re-screens
  • Browsing top-ranked agents

Pricing

Freemium, from $12 a month

Free for public scores; Pro $12/mo, Team $29/mo, Enterprise from $499/mo

Prices change. Check badgeia.com before you buy.

Badge MCP Server FAQ

Is it free?

Public scoring is free; Pro is $12 a month, Team $29 a month, and Enterprise starts at $499 a month billed annually.

Do I need a public endpoint?

Yes for a verified score. Without one the run is simulated and cannot earn a certificate.

What does verified mean?

The run was recorded by Badge and has not been altered; it does not claim the task was hard.

Does Badge see my code?

No. It calls your endpoint like any client and records results.

How many runs are included?

Free allows 30 per month, Pro 1,000 and Team 5,000.

Is Badge MCP Server free?

Badge MCP Server has a free plan, with paid plans for more usage or features. Paid plans start at $12 a month.

How do I add Badge MCP Server to Claude Code?

Run this command in your terminal: claude mcp add badge-mcp-server --env BADGE_API_KEY='<BADGE_API_KEY>' -- npx -y @badgeia/mcp-server

What are the best alternatives to Badge MCP Server?

The closest alternatives to Badge MCP Server are Claude Code Ultimate Guide, signafAIde, Toolkien, Lugh and HypePaper.

Alternatives to Badge MCP Server

  1. MCP server for searching the Claude Code Ultimate Guide from any MCP client

    Runs locally
    • Open source
    • 6.5K GitHub stars
    • AI models and LLM tooling
  2. signafAIde

    MCP server

    Local MCP server to set up, verify and check coverage of signed agent action records

    Runs locally
    • 5 tools
    • Freemium
    • Security
  3. Toolkien

    MCP server

    One MCP gateway for all your MCP servers, with code mode, permissions and audit

    Not checked yet
    • 3 tools
    • Remote
    • Open source
    • AI models and LLM tooling
  4. Lugh

    MCP server

    Write, check and dry-run Lugh agent programs with policy, budgets and approvals

    Runs locally
    • 5 tools
    • Open source
    • AI models and LLM tooling
  5. HypePaper

    MCP server

    Research papers, benchmarks, citations and topic experts for AI assistants

    • 58 tools
    • Remote
    • Freemium
    • AI models and LLM tooling