Skip to main content
Jam is built to be used from an agent. Everything in the dashboard is a tool on the Jam MCP server and a route on the REST API, so your agent can set Jam up, read where you stand, and do the work on the Plan.

Pick your client

Claude

Claude.ai and Claude Desktop. Add a custom connector and sign in.

Claude Code

One command. Sign in in the browser, or use an API key.

ChatGPT

Add a connector in developer mode and sign in.

Cursor

Add the server to mcp.json with an API key.

Other MCP clients

Any client that speaks Streamable HTTP, with OAuth or a key.

No account yet

Your agent sets Jam up for you and sends you one link to approve.

Two ways to sign in

OAuth (no key to paste). The client opens a Jam page in your browser. You sign in (or sign up and choose a plan), check the workspace and where you will be sent back to, and approve. Jam supports OAuth 2.1 with dynamic client registration and PKCE, so any MCP client that can sign in to a remote server works this way, as long as its sign-in returns to an https address or to localhost. API key. Create one in Settings, Developer, Create API key. It is shown once. Keys look like sk_ followed by 64 hex characters. The client sends it on every request as Authorization: Bearer sk_.... Either way the connection belongs to one workspace and acts with your role in it. An OAuth sign-in shows up in Settings, Developer as a key named after the client, for example “Claude (connector)”. Revoke it there to disconnect.

Then what

Once connected, ask your agent something like “How visible are we in AI answers, and what should we do first?” See What your agent can do for the order of work, and the operating guide for instructions to paste into your agent’s context. Already using the dashboard? Every action has a Copy prompt button that hands one job to any agent, connected or not. See Copy prompt.