Free
- Tools at /mcp from your routes
- Agents sign in through your own login
- Hosted broker, up to 5 agent connections per project
- 1 project
- A small "Powered by agemno" line on your Connect page
- Tool-call logs Coming soon, 7-day retention
One line makes your app an MCP server. Reads go live on their own. Writes wait until you opt them in.
$ npx agemnoSetup reads your app and lists what an agent would get. Reads go on by themselves. Writes stay hidden until you opt them in.
agemno reads your Markdown, splits it at every heading, and gives agents a search_docs tool next to your routes. Every file is also an MCP resource they can open whole.
README.md, docs/**/*.md by defaultsearch_docs, read-only, plus agemno://docs/…app.use(agemno({ docs: ["README.md", "docs/**/*.md"] })); // the default
Live An LLM answers using agemno's real docs below — the same search_docs tool your app gets — gated by a retrieval-score check and an on-topic check before it ever runs.
This is agemno's own docs: 9 files, 102 sections, split by the same rules and searched with the same MiniSearch settings as search_docs, right in your browser.
Agents gather from wherever your users work.
Claude, ChatGPT, Cursor and VS Code, the clients your users already run.
Every call lands at /mcp, as the signed-in user.
Every call runs as the user.
Through your app's own middleware, with exactly the caller's credentials.
Billed monthly through Polar. No card to start on Free.
agemno.
0000
The real thing needs no aim. Reads are exposed by themselves; a write answers only once you wrap it in exposed().
Enter to play again · Esc to leave