AgentoolRank MCP server
Search and compare open-source AI agent tools from your AI assistant, ranked by live GitHub activity.
AgentoolRank tracks hundreds of open-source AI agent tools (frameworks, coding agents, memory, RAG, evals, MCP servers) and ranks them by GitHub activity: stars, star pace, commits, releases and package downloads, refreshed daily. This repository documents how to connect to its hosted MCP server and JSON API. The server runs on agentoolrank.com; there is nothing to install or self-host.
- MCP endpoint:
https://agentoolrank.com/api/mcp(Streamable HTTP, no auth required) - Official MCP Registry:
com.agentoolrank/agent-tools - Optional API key: one click, no signup, at agentoolrank.com/api-key. Send it as
Authorization: Bearer <key>to identify your agent.
Tools
| Tool | What it does |
|---|---|
search_tools(query, limit?) | Find tools for a job, ranked by relevance then GitHub activity. |
get_tool(slug) | Stars, 30-day star growth, capabilities, limitations and who it is best for. |
get_alternatives(slug) | Open-source alternatives to a tool, with their GitHub stats. |
submit_tool(url, name, tagline, email, …) | List a tool. The response shows every pricing option up front, including the free one. |
get_submission_status(submission_id, status_token) | Review status of a submission. |
recommend_directories(product_type) | Launch directories our own products went through, with free-tier conditions and the steps only a person can do. The top 5 for each product type are free. |
next_directory(key, product_type?) | Submit Kit key (9): hands your agent the next directory that fits the product, one at a time (up to 50, at most 3 unreported). |
report_result(key, domain, status, listing_url?) | Submit Kit key: report what happened at a directory. Reports are leads; listing rates count only listings our own daily check verifies. |
my_progress(key) | Submit Kit key: every directory handed out so far and its reported status (also at agentoolrank.com/submit-kit/progress). |
Prompts
Ready-made prompts your MCP client can offer (prompts/list, prompts/get):
| Prompt | Arguments | What it does |
|---|---|---|
compare_tools | tool_a, tool_b | Compares two tools side by side: GitHub stars, 30-day star growth, last release, downloads, best for, limitations |
find_alternatives | tool | Lists open-source alternatives to a tool, ranked by live GitHub activity |
launch_plan | product_type (ai_tool, mcp_server, dev_tool, saas, other), optional product_url | Plans where to list a new product: which launch directories fit it and which to skip |
Resources
| URI | Content |
|---|---|
agentoolrank://about | How the ranking works and what this server offers (Markdown) |
agentoolrank://top-tools | The current top 20 open-source AI agent tools by GitHub activity, live (Markdown table) |
Connect
Claude Code
claude mcp add --transport http agentoolrank https://agentoolrank.com/api/mcp
Cursor
Add to ~/.cursor/mcp.json (or .cursor/mcp.json in a project). See examples/cursor-mcp.json:
{
"mcpServers": {
"agentoolrank": { "url": "https://agentoolrank.com/api/mcp" }
}
}
Claude Desktop
Add it under Settings → Connectors as a custom connector with the URL above. Or use the mcp-remote bridge in claude_desktop_config.json. See examples/claude_desktop_config.json.
Codex CLI
Add to ~/.codex/config.toml. See examples/codex-config.toml.
JSON API
The same data is also available without MCP. See examples/curl.sh.
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
About
AgentoolRank is run by TENSO LLC. Questions or corrections: hello@agentoolrank.com. Privacy · Terms.
The contents of this repository (documentation and configuration examples) are MIT licensed. The AgentoolRank website, data and service are not part of this repository.