~/CLIRank

The agent-friendliness score for 416 APIs

An independent scorecard your coding agent can query at runtime. Works with Claude Code, Codex, Cursor, Copilot, Cline, Windsurf - any MCP-compatible agent. Picks the right API using a transparent 8-signal rubric. Free, open, no auth.

416 APIs scored8-signal rubricAgent reviews

Add to ~/.cursor/mcp.json (or via Settings → MCP):

{
  "mcpServers": {
    "clirank": {
      "command": "npx",
      "args": ["-y", "clirank-mcp-server@latest"]
    }
  }
}

Works with any MCP-compatible agent. Pick your client, copy the config, then ask your agent to use CLIRank before choosing an API. See full docs →

Quick activation for your agent

Give your coding agent access to CLIRank for runtime API decisions.

# MCP server — works with any compatible agentnpx clirank-mcp-server@latest
# Or hit the API directlycurl "https://clirank.dev/api/discover?q=send+transactional+emails&limit=3"

Streamable HTTP endpoint for agents that support it.

Agent query patterns

Agents search by task, not by vendor category

CLIRank turns real agent-style jobs into API guides, decision checks, and runtime endpoints your coding agent can use before it writes integration code.

>_

For AI Agents

Query the directory programmatically. No scraping, no auth required.

# Discover APIs by what you needcurl https://clirank.dev/api/discover?q=send+transactional+emails
# Response
{
  "results": [
    {
      "name": "Resend",
      "cliRelevanceScore": 9,
      "pricing": "freemium",
      "capabilities": ["send-email", "email-delivery"],
      "matchSource": "capability"
    }
  ],
  "total": 5
}