Describe the app. Watch it get built.
Studio is a coding team you talk to. Ask for a feature in plain language and the agents write the code on your own Foreman host, run it, and render a live preview right in the app — while you watch.
Let it build — or just think out loud.
Every project has a chat. Pick how much rope the agent gets, per message.
Full permissions. The agent writes the code, runs it, and verifies it before calling it done.
Read-only. It investigates the codebase and proposes a plan — nothing changes until you say so.
Just talk it through. No tools, no edits — think out loud with an agent that knows your code.
A running app, not a wall of diffs.
Studio renders your project's dev server live in a Preview tab, streams the real logs next to it, and gives you the file tree and an editor when you want to look under the hood. Set environment variables per project — secrets are encrypted at rest and written straight to your .env.
Connect a repo. Deploy in a click.
Link a GitHub, GitLab, or Bitbucket repo — no tokens to paste — and commit and push from inside Studio. When it's ready, one click deploys to your connected Render or Vercel account. And before the agent ever claims it's done, it runs a typecheck, a build, and a headless smoke test against the live app.
Your code. Your host. Your models.
Code, dev server, and files live on your Foreman host — the desktop app or a cloud host — not in a black box.
Scaffold a fresh project from a template, or open a codebase already on your machine and pick up where it left off.
Choose which AI models the coding agents use with a workspace-scoped allowlist — separate from the rest of your team.
Studio runs on your Foreman host — the desktop app or a cloud host. When no host is online, previews and the dev server pause until you bring one back up.
Build your next thing by describing it.
Free to start — no card needed. Open Studio, say what you want, and watch it come together on your own machine.