AI coding

Ask in plain English and the agents write and run Python, SQL, or R for you — with every artifact open to read, verify, and re-run.

You don’t have to write code — but when code is the right tool, the agents write and run it for you.

Plain English → running code

Describe what you want and the agent generates Python, SQL, or R, runs it on the server-side runtimes, and renders the result. It’s the same path you’d take with the editor and terminal yourself — just driven by conversation.

Transparent and editable

Whatever the agent writes appears as an artifact you can open in the editor to read, verify, tweak, and re-run. Nothing is hidden, and results are deterministic and repeatable.

Hand off either way

Start in plain English and drop into code for a precise change, or write code and ask the agent to extend it. You and the agents share the same workspace, so control passes either direction at any time.