Skip to main content

Create a dashboard with AI

The AI that builds a Vibedasher dashboard is your agent, not ours. It runs in your own harness — Claude Code, Codex, Cursor, any MCP client — it writes the React and the SQL locally, and it reaches Vibedasher through the MCP server. We maintain the half that should not be rebuilt in every session: connections, datasets, pipelines, refreshes, governance, the query engine, hosting.

That is a deliberate split rather than a limitation. Your agent is already good at writing a React chart; what it cannot do in a one-off session is keep a data platform alive between sessions.

Start here​

pip install vibedasher-mcp
vibedasher login
claude mcp add vibedasher --env VIBEDASHER_REGION=eu-central-1 -- vibedasher-mcp

Then ask your agent for the dashboard you want, in the same words you would have typed into a chat box. It will get the data in, explore it with SQL, write the tiles, build, verify, and publish — and hand you a link.

If you would rather not run an agent yourself​

The app at cloud.vibedasher.com is where you browse, open, publish and share dashboards, manage datasets and connections, and set up who can see what — all of it on the same objects your agent works on. You can mix freely: an agent builds, a person reviews in the app, either one publishes.

What the app is not is the place we would point you for building a dashboard. The terminal path above is the one we build for, document, and test.

Next step​

Once your agent has published something, the code is still yours: eject the dashboard and host it wherever you like.