Developers
Connect AI assistants and your own tools to Plavn with personal access tokens and the Model Context Protocol (MCP) server. Start here for setup, addresses and security.
Bu içerik henüz dilinizde mevcut değil.
Plavn can be used from AI assistants and other tools you already work with. Today that means one thing: an MCP server, through which an assistant such as Claude Code, Claude Desktop or Cursor can look things up in your tasks and mail and, if you allow it, add tasks and ask Plavn to prepare replies.
There is no public REST API yet. A personal access token works with the MCP endpoint only.
Where to start
Section titled “Where to start”| Page | What it covers |
|---|---|
| MCP server | What the server offers and a five-minute setup. |
| Personal access tokens | Creating, naming, limiting and revoking tokens. |
| Connect a client | The endpoint, ready-made settings for common clients, a local (stdio) setup for self-hosting, and fixing errors. |
| Tools reference | Every tool, its inputs, limits and what it returns. |
| Security and privacy | What an assistant can see, how to limit it, and mail that tries to steer it. |
Addresses
Section titled “Addresses”| Address | Purpose |
|---|---|
app.plavn.com |
The signed-in app. |
api.plavn.com |
The API, including the MCP endpoint at /mcp. |
docs.plavn.com |
This documentation. |
If you run Plavn yourself, the endpoint is your own backend’s address followed by /mcp. Settings → Developer always
shows the exact address for the Plavn you are using.