AgentSearch

AgentSearch

v0.1.0• Luther Systems

Find AI agents and MCP servers by natural-language query. AgentSearch indexes ~3,700 hosted agents daily, probes each for reachability, scores them on reputation + usability + functionality, and returns the best matches for your task with per-agent reasoning. Use it to answer 'is there an agent that does X?' or 'is this specific agent any good?'. Read-only, free, no auth.

Launch agent website ↗

Skills

• Search the agent index

  Take a natural-language query and return the top ranked agents/MCP servers from the index, with per-agent reasoning. Ranked by final_score = query_fit × agent_quality.

  searchdiscoveryrankingagent-discoverymcp-discovery

• Get a scored agent profile

  Return the full scored profile for one named indexed agent: inventory data, layer-by-layer scores (Fame / Usability / Functionality / Call Graph), reasoning, and live blocker status.

  detailsprofilescoring

• Live-score an arbitrary agent URL

  Crawl + probe + LLM-score an arbitrary agent URL using the same rubric the batch pipeline uses. Works for agents not (yet) in the index.

  scoringliveprobeagent-evaluation

• Browse the full index

  Paginated list of every indexed agent, ranked by overall quality.

  browselistrank

• Index summary stats

  Distributions, top categories, and the agent-to-agent call graph for the whole index.

  statsrollupsgraph

Integration

import asyncio
from a2a_registry import AsyncRegistry

async def main():
    async with AsyncRegistry() as registry:
        agent = await registry.get_by_id("2922b3a6-6f7f-42f4-98ee-2ef52bb46255")
        client = await agent.async_connect()
        print(f"Connected to {agent.name}")

asyncio.run(main())