Google PageRank for AI agents. 25,000+ tools indexed.
1

Search the index

The search endpoint is public — no key needed. Query for any tool, skill, or agent topic:

curl
curl "https://agentrank-ai.com/api/v1/search?q=filesystem&limit=3"
Response
{
  "query": "filesystem",
  "category": "all",
  "sort": "score",
  "results": [
    {
      "type": "tool",
      "id": "modelcontextprotocol--servers",
      "name": "modelcontextprotocol/servers",
      "description": "Model Context Protocol servers",
      "score": 97.4,
      "rank": 1,
      "url": "https://agentrank-ai.com/tool/modelcontextprotocol--servers/"
    }
  ],
  "meta": { "total": 48, "limit": 3, "offset": 0 }
}
2

Get a specific tool

Look up any indexed tool by its GitHub owner/repo. Replace / with -- in the URL:

curl
curl "https://agentrank-ai.com/api/v1/tool/modelcontextprotocol--servers"

The response includes the score, rank, signals breakdown, and raw GitHub data.

3

Use the SDK (optional)

Install the TypeScript client to skip manual URL construction:

npm
npm install @agentrank/sdk
TS
import { AgentRank } from '@agentrank/sdk';

const ar = new AgentRank(); // no key = free tier

const results = await ar.search('filesystem', { limit: 5 });
console.log(results.results[0].name, results.results[0].score);

const tool = await ar.getTool('modelcontextprotocol/servers');
console.log(tool.rank, tool.signals);
4

Get an API key for more

The free tier gives you 100 req/min per IP with no key. A Pro key unlocks:

  • 10,000 requests/day
  • Rank history on every tool — 30-day trend
  • Batch lookup: resolve 50 tools in one request
  • Access to v2 endpoints with additional fields

Get a key at agentrank-ai.com/pricing/. Keys look like ark_pro_...

TS
import { AgentRank } from '@agentrank/sdk';

const ar = new AgentRank({ apiKey: 'ark_pro_your_key_here' });

// Now using v2 — includes rankHistory
const tool = await ar.getTool('modelcontextprotocol/servers');
console.log(tool.rankHistory); // 30-day rank trend

Next steps