Agenlify MCP server
Let your AI assistant find tools for a task, inspect their documented capabilities and explain the tradeoffs with sources.
Connect your assistant
Add a remote MCP server in a client that supports Streamable HTTP. Use the endpoint below. No Agenlify account or API key is required.
https://agenlify.ai/mcpFor clients that accept an mcpServers JSON configuration with HTTP transport:
{
"mcpServers": {
"agenlify": {
"type": "http",
"url": "https://agenlify.ai/mcp"
}
}
}Client configuration formats vary. If your client has an “Add remote server” screen, enter the endpoint URL there instead.
Three tools, one public catalog
The server reads the same 184 product listings and 11 published task guides as the website. It returns public catalog data and does not call the listed products.
search_agents— Find products by English or Chinese task, product name or official domain. Filter by task, category, pricing, setup or evidence. Use limit and offset to page through results, up to 20 at a time.get_agent— Read one product by slug, including setup steps, pricing context, limitations, review dates, official sources and recorded cases.list_tasks— Discover supported task IDs, preparation checklists, expected outputs and review requirements.
Try a concrete request
“Find tools for analyzing a spreadsheet. Compare the inputs, setup and pricing of three candidates, then show the source links and what I need to check myself.”
“Look up Cursor in Agenlify. Explain its documented workflow, limitations and pricing review date.”
What the results mean
Search uses the directory’s controlled task vocabulary and name matching. It is not semantic search and can return no match. Results are ordered by name and task relevance, with alphabetical ties; this is not a quality ranking.
Official-source review is not hands-on testing. The tested filter only identifies products with a separately recorded case, not a general performance guarantee. Product details and prices can change; follow the returned sources and check the review dates. Free access can have limits or separate model costs.
Access and privacy
The MCP endpoint is free to access. Your AI client or model provider may charge separately. All three tools are read-only: they do not access saved products, accounts or private user data, and do not run vendor tools or submit forms.
Calls are stateless. The endpoint does not save MCP queries or tool arguments to the application database. Hosting security and operational logs may record request metadata. See our privacy policy for the website’s separate account and analytics practices.
Native MCP clients can connect directly. Cross-origin browser requests are not supported. Request bodies are limited to 16 KiB; search queries to 300 characters.
Explore and verify
Browse the public directory · Read task guides · How we verify products · Privacy policy