Every side project I start, I rebuild the same backend. Auth, roles, file uploads, emails. So I stopped rebuilding it. I built Feathears BaaS instead ! An open-source CLI that scaffolds a production-ready Node.js backend in one command. Auth with argon2id + JWT refresh, RBAC, file uploads, email notifications, database migrations, all wired up and ready to deploy in one command. From First command to First API call in five minutes. But the real reason it exists is different. I wanted a backend that LLMs could actually extend without breaking things. Every Feathers BaaS generated project is self descriptive. Run `feathers-baas describe` and the agent sees your entire app as JSON: every service, method, hook, and migration. Generated projects also ship with an AGENTS.md that tells Claude Code, Codex or Cursor exactly how to add a new resource without guessing. The result: you can hand the project to an agent and say "add a posts service with title and body" and it produces architecturally consistent code on the first try. Not because the model is smarter. Because the codebase is predictable enough that there are fewer ways to be wrong. Finally, Feathers BaaS is: self-hosted and database-agnostic ( support PostgreSQL, MySQL, SQLite and MongoDB out-of-the-box). Plug-in your storage (S3, GCS or local), your email provider (currently support any SMTP provider, Resend, Brevo and Sendgrid). You are not locked-in: the CLI is a scaffolder, not a runtime dependency. What gets generated is yours. Run the command below and watch your agent do the rest. npx feathers-baas init my-api For further reading: https://feathersbaas.com/ https://github.com/theComicsdeal...

Feathers BaaS

Production-ready backend in one command. Built for agents.

Production-ready backend in one command. Built for agents.