AgentoolRank for AI agents
Your agent can search and compare 621 open-source AI agent tools (frameworks, coding agents, MCP servers, RAG, evals…) ranked by live GitHub activity, and list new tools, without a browser.
MCP server
Streamable HTTP, no auth required. Also listed in the official MCP Registry as com.agentoolrank/agent-tools. Optional: a free API key (one click, no signup), sent as Authorization: Bearer <key>, identifies your agent.
{
"mcpServers": {
"agentoolrank": { "url": "https://agentoolrank.com/api/mcp" }
}
}search_tools(query, limit?)— find tools for a job, ranked by relevance then GitHub activityget_tool(slug)— stars, 30-day growth, capabilities, limitations, best forget_alternatives(slug)— open-source alternatives with their statssubmit_tool(url, name, tagline, email, …)— list a tool; returns every pricing option up frontget_submission_status(submission_id, status_token)— review status
claude mcp add --transport http agentoolrank https://agentoolrank.com/api/mcp
Prompts
compare_tools(tool_a, tool_b)— side-by-side table: stars, 30-day growth, last release, downloads, best for, limitationsfind_alternatives(tool)— open-source alternatives ranked by live GitHub activitylaunch_plan(product_type, product_url?)— which launch directories fit a new product and which to skip
Resources
agentoolrank://about— how the ranking worksagentoolrank://top-tools— the current top 20, live
JSON API
GET https://agentoolrank.com/api/v1/tools?q=rag&limit=10
GET https://agentoolrank.com/api/v1/tools/{slug}
POST https://agentoolrank.com/api/v1/submissions
{"url": "...", "name": "...", "tagline": "...", "email": "...",
"github_url": "...", "max_budget_usd": 20, "deadline_days": 3}Choosing other launch directories
recommend_directories(product_type) returns launch directories our own products actually went through, tagged auto or manual, with free-tier conditions, the link type we measured, form tips, the confirmation to look for, and one checklist of the steps only a person can do. Free: top 5. With a Submit Kitkey: up to 50, one at a time sites plus a don’t-submit list with reasons. It does not submit for you and never bypasses captchas.
Listing a tool on behalf of your human
- Free listing is reviewed in queue order; showing our badge on the product site moves it up.
- The response lists all options at once: free, $9 (live within 3 days), $19 (1 day), $49 (1 day + 7 days featured on the homepage).
message_for_humanis a short summary of the result and the optional upgrades, written so your agent can forward it to you as-is.- Pass
max_budget_usdanddeadline_daysand we return the cheapest option that fits asrecommended_plan. - Paid options come with a
checkout_urlfor your human to pay. No upsells later; full refund if not approved.
Machine-readable summary: /llms.txt
MCP server FAQ
- What is the AgentoolRank MCP server?
- The AgentoolRank MCP server is a remote server using Streamable HTTP at https://agentoolrank.com/api/mcp. It lets AI assistants search and compare open-source AI agent tools ranked by live GitHub activity.
- How do I add it to Claude Code, Cursor or other MCP clients?
- For Claude Code, run: claude mcp add --transport http agentoolrank https://agentoolrank.com/api/mcp. For Cursor or other clients that support Streamable HTTP and mcpServers configuration, use: { "mcpServers": { "agentoolrank": { "url": "https://agentoolrank.com/api/mcp" } } }
- Do I need an API key?
- No API key is required to connect, search or compare tools. An optional free API key from https://agentoolrank.com/api-key, sent as Authorization: Bearer YOUR_API_KEY, identifies your agent. Paid Submit Kit tools require a separate Submit Kit key.
- How are tools ranked and how fresh is the data?
- AgentoolRank ranks open-source AI agent tools using GitHub stars, star growth, commits, releases, and npm/PyPI downloads. Data is refreshed daily. Search results are ranked by relevance, then GitHub activity.
- Can I list my own MCP server or agent tool?
- Yes. Use submit_tool via MCP or https://agentoolrank.com/submit. The free review queue processes 5 submissions per day. Priority review costs $9 with review within 24 hours; fast-track review costs $19 with review within 2 hours. These are review timelines, not guarantees of approval.
- Which launch directories should I submit my MCP server to?
- Call recommend_directories with product_type "mcp_server" for the top 5 matching directories for free, including submission conditions and human steps. The Submit Kit costs $29 once and offers up to 50 matching directories, handed out one at a time via next_directory as you report results with report_result. The key is valid for 90 days. Get it at https://agentoolrank.com/submit-kit. The kit provides guidance; it does not submit for you.