CrescoDB

Most backends start simple and end up as a bill. You pick a hosted database, bolt on auth, storage and realtime, and somewhere along the way the thing you built lives on someone else's servers and answers to their pricing page.

CrescoDB takes the other route. One command gives you a working backend, and you deploy it to a server you rent.

From a single schema you get a database on SQLite, Postgres or MySQL, an auto-generated REST API, auth with roles and row-level security, file storage and realtime. Build locally on SQLite and ship to Postgres or MySQL in production with no rewrite. Deploy over SSH to your own box, or through Render or Railway, with HTTPS certificates, one-click rollback and Git auto-deploy.

Then it keeps working after the deploy. CrescoDB watches your app from the inside and the outside, alerts that name what broke, and takes a nightly backup. Every backup gets restored into a scratch database on your own server before it is trusted, then restored again on the way back. You find out a backup is broken before you need it, not after.

Nothing is a hostage. cresco eject writes the code and every row out as plain SQL. If CrescoDB ever stops operating, the CLI and engine source is published under Apache 2.0.

cresco mcp lets Claude Code, Cursor or Codex work on a CrescoDB project directly, and an AI assistant drafts the schema and writes real, editable code.

Free to build locally. Pro is $20 a month with nightly verified backups, alerts and 15 apps across 5 servers. Team is $45 per seat and Enterprise $185, billed per account rather than per project.

For builders who want to own what they ship.