AgentRank

Google PageRank for AI agents. 25,000+ tools indexed.

Integrations

Connect your client to AgentRank's hosted Streamable HTTP MCP endpoint, or use the npm package when your client requires a local stdio process.

How it works

Hosted (recommended): connect to https://agentrank-ai.com/mcp; no local AgentRank process or API key is required for the public tools. Local fallback: clients that cannot use Streamable HTTP can launch npx -y agentrank-mcp-server over stdio. Adding either connection only makes the tools available: your client, instructions, and approval settings determine whether and when they are called.

Claude Code

Hosted MCP

Hosted Streamable HTTP (recommended)

Run claude mcp list or open /mcp in Claude Code to verify the connection.

claude mcp add --transport http agentrank https://agentrank-ai.com/mcp
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

claude mcp add agentrank -- npx -y agentrank-mcp-server

Official Claude Code MCP documentation

Cursor

Hosted MCP

Hosted Streamable HTTP (recommended)

Merge this entry into .cursor/mcp.json (project) or ~/.cursor/mcp.json (user). Preserve existing servers.

{
  "mcpServers": {
    "agentrank": { "url": "https://agentrank-ai.com/mcp" }
  }
}
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

{
  "mcpServers": {
    "agentrank": { "command": "npx", "args": ["-y", "agentrank-mcp-server"] }
  }
}

Official Cursor MCP documentation

VS Code + GitHub Copilot

Hosted MCP

Hosted Streamable HTTP (recommended)

Merge this server into .vscode/mcp.json or your VS Code user-profile mcp.json, then start it from MCP: List Servers.

{
  "servers": {
    "agentrank": { "type": "http", "url": "https://agentrank-ai.com/mcp" }
  }
}
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

{
  "servers": {
    "agentrank": { "type": "stdio", "command": "npx", "args": ["-y", "agentrank-mcp-server"] }
  }
}

Official VS Code + GitHub Copilot MCP documentation

Cline

Hosted MCP

Hosted Streamable HTTP (recommended)

In Cline, open MCP Servers → Remote Servers, enter the URL, and choose Streamable HTTP. Keep tool approval decisions under your control.

{
  "mcpServers": {
    "agentrank": {
      "type": "streamableHttp", "url": "https://agentrank-ai.com/mcp",
      "disabled": false, "autoApprove": []
    }
  }
}
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

{
  "mcpServers": {
    "agentrank": {
      "command": "npx", "args": ["-y", "agentrank-mcp-server"],
      "disabled": false, "autoApprove": []
    }
  }
}

Official Cline MCP documentation

Windsurf

Hosted MCP

Hosted Streamable HTTP (recommended)

Open Cascade → MCPs → Add Server, select Streamable HTTP, and paste this endpoint. Windsurf stores MCP settings in ~/.codeium/windsurf/mcp_config.json.

https://agentrank-ai.com/mcp
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

{
  "mcpServers": {
    "agentrank": { "command": "npx", "args": ["-y", "agentrank-mcp-server"] }
  }
}

Official Windsurf MCP documentation

OpenAI Codex

Hosted MCP

Hosted Streamable HTTP (recommended)

Run codex mcp list to verify. Codex CLI and the Codex IDE extension share the same MCP configuration.

codex mcp add agentrank --url https://agentrank-ai.com/mcp
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

codex mcp add agentrank -- npx -y agentrank-mcp-server

Official OpenAI Codex MCP documentation

OpenCode

Hosted MCP

Hosted Streamable HTTP (recommended)

Merge this into opencode.json (project) or the global OpenCode config, then run opencode mcp list.

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "agentrank": { "type": "remote", "url": "https://agentrank-ai.com/mcp", "enabled": true }
  }
}
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "agentrank": { "type": "local", "command": ["npx", "-y", "agentrank-mcp-server"], "enabled": true }
  }
}

Official OpenCode MCP documentation

OpenClaw

Hosted MCP

Hosted Streamable HTTP (recommended)

Run openclaw mcp doctor agentrank --probe for a live tool-discovery check.

openclaw mcp add agentrank --url https://agentrank-ai.com/mcp --transport streamable-http
Show npm/stdio fallback

Use this only when hosted Streamable HTTP is unavailable. It requires Node.js and launches the published npm package locally.

openclaw mcp add agentrank --command npx --arg -y --arg agentrank-mcp-server

Official OpenClaw MCP documentation

Available tools

Once connected, your client can discover five read-only advisory tools:

search

Search 25,000+ MCP servers, agent tools, and skills by keyword. Returns scored results ranked by real GitHub signals.

e.g. "browser automation", "database", "slack"

lookup

Check if a specific GitHub repository is in the index. Returns score, rank, and signal breakdown.

e.g. github.com/microsoft/playwright-mcp

evaluate

Evaluate a tool against AgentRank's maintenance, adoption, and quality signals.

recommend

Get ranked recommendations for a use case and optional requirements.

alternatives

Find ranked alternatives to a named tool or repository.

API quickstart

Prefer direct API calls? The REST API is free to use, no key required. Free tier: 100 requests/min per IP. Pro keys unlock 10,000 req/day.

GET /api/v1/search Search tools by keyword
curl
curl "https://agentrank-ai.com/api/v1/search?q=browser+automation&limit=5"
Response
{
  "query": "browser automation",
  "results": [
    {
      "type": "tool",
      "id": "microsoft--playwright-mcp",
      "name": "microsoft/playwright-mcp",
      "description": "Playwright MCP server for browser automation",
      "score": 94.2,
      "rank": 3,
      "url": "https://agentrank-ai.com/tool/microsoft--playwright-mcp/"
    }
  ],
  "meta": { "total": 127, "limit": 5, "offset": 0 }
}
GET /api/v1/tool/:slug Get a specific tool's score and signals
curl
curl "https://agentrank-ai.com/api/v1/tool/microsoft--playwright-mcp"
Response
{
  "slug": "microsoft--playwright-mcp",
  "name": "microsoft/playwright-mcp",
  "score": 94.2,
  "rank": 3,
  "signals": {
    "stars": 8200,
    "starsScore": 22,
    "freshnessScore": 25,
    "issueHealthScore": 20,
    "contributorsScore": 10,
    "dependentsScore": 17.2
  },
  "lastCommitDate": "2026-03-17"
}
GET /api/v1/top Top-ranked tools (default: top 20)
curl
curl "https://agentrank-ai.com/api/v1/top?limit=10&category=tools"

Rate limit headers: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset on every response. See rate limit docs for details.

API docs • All configs on GitHub • npm package