Quick start
Most people never touch a terminal. You tell your AI assistant to publish, and it does the rest.
Percher is built to be driven by your AI assistant: Claude, Cursor, Windsurf, or whatever you build with. You don't install anything or learn any commands. You ask it to publish, in plain words.
Publish my app to percher.app.
Paste this to Claude, Cursor, or any AI assistant.
Changed something and want the live version updated? Say “Publish my latest changes to percher.app.” It's the same flow.
What happens next
- 1The first time, you sign in.Your assistant opens a browser window (or hands you a link). Click approve. That's a one-time step, and after that it just works.
- 2It sets everything up for you. No config files to write, no server to pick. Your assistant detects what kind of app you built and handles the details.
- 3You get a link back. A couple of minutes later your assistant hands you a real web address like
your-app.percher.run, already secured with HTTPS. Share it with anyone.
That's the whole thing. Your app is live at your-app.percher.run with SSL, builds, and a health check, typically in about 2 minutes.
For agents and developers
Prefer to run it yourself? Percher also ships as a CLI, with no global install needed. Run it with bunx (Bun) or npx (Node). One command does login, config, and deploy:
# One command does everything: login, config, deploy cd my-app bunx percher publish # Granular control (only when you need to isolate a single phase): bunx percher login bunx percher init bunx percher push