Start
Stableship docs
Set up Stableship, connect your coding agent, and keep every release of your app checked.
What Stableship does
Stableship is QA and security for apps built with AI, on GitHub and Supabase. It reads your code and your database structure, tests a private copy of your app as every kind of user, and tells you what broke and who it affects. Your coding agent writes the fix, and Stableship checks it. Stableship never writes to your repository.
Start here
- New to Stableship? Follow the Quickstart: connect GitHub and Supabase, confirm who should see what, and start your first audit. It takes about ten minutes.
- Want the ideas first? Read How Stableship works: the private copy, audits and checks, kinds of user, findings and verdicts.
- Something not working? The Help pages answer the questions people ask most, and Getting help says how to reach us.
What you need
- An app with its code on GitHub.
- Its database on Supabase, including Lovable Cloud and Bolt's; see Connecting Supabase.
- A coding agent, such as Claude Code, Cursor, Codex, Windsurf, Lovable, Bolt, Replit or v0. It is optional, but it is how fixes get made.
For coding agents
Every page in these docs is also plain Markdown: add .md to its address, for example /docs/quickstart.md. The Stableship MCP server is at https://api.stableship.ai/mcp, and a short summary for agents is at /llms.txt.
For your coding agent: this page as Markdown