A board that follows your workflow.

Give your AI client access to the board.

Kardbrd exposes board tools through Model Context Protocol. Connect a compatible client to work with the cards and boards your account can access.

Add the Kardbrd server.

In your client’s remote MCP settings, add this server URL:

https://app.kardbrd.com/mcp/

Complete the authorization flow in your browser. Use your own account and confirm the workspace access. Client setup screens and supported transports vary; follow your client’s maintained instructions.

Check the connection with a read.

Ask the client to list your available boards or read a card you identify. Confirm it is using the intended account and board before asking it to change anything.

For clients that use a JSON server configuration, the endpoint is the same; the configuration keys depend on the client. Avoid placing access tokens in shared source files.

Know what the connection supplies.

MCP provides access to board operations such as reading and updating tasks. The client still needs its own tools and permissions to work on an application, run checks, or release changes.

For a daemon that responds to configured board events and invokes an executor, use the Kardbrd agent.

Include the client in your data review.

The client and its model provider may process the board material you ask it to use. Review the access granted and the provider’s terms alongside your operating setup.

Read the privacy policy and the working principles for responsible use.

What work do you want to get moving?

Tell us what you want to get done and who handles it today. We can help you use the board with your own team, or discuss having our team do the work.

Talk to our team