Is CrescoDB a database?

No. It uses a database. CrescoDB runs on SQLite, PostgreSQL or MySQL and generates the API, auth and dashboard around whichever one you point it at.

Do I need an account to use it?

No. Local development needs no account, no card and no cloud project. You install the CLI and it runs.

Do I have to write Node?

No. Your app talks to the generated REST API over HTTP from any language. You only need Node installed to run the CLI.

What is CrescoDB?

Short answer: a backend toolkit you run yourself. Not a service you rent, and not a database with a dashboard bolted on.

The one-line version

You describe your data in one file. You get a database, a full REST API, auth with roles, file storage and realtime. One command, on your own machine, with no account and no card.

$ npx crescodb init $ cresco dev REST API on :3000 · dashboard ready

What you actually get

An auto-generated REST API. Full CRUD, filtering, ordering, pagination and bulk insert for every model. You do not write a route.
Auth with roles and per-row rules, so one user can never read another user's rows.
The same schema on SQLite, PostgreSQL and MySQL. Build on SQLite with zero config, ship on Postgres, and there is no rewrite waiting at deploy.
File storage and realtime, with signed URLs and live subscriptions over SSE.
A dashboard with a data browser, a SQL editor and a request explorer that already found your endpoints.
Deployment to a server you own, in one command, with automatic HTTPS.

How it differs from a hosted backend

A hosted backend gives you a URL and keeps the machine. That is genuinely convenient, and it comes with three things people complain about consistently: projects that pause, bills that surprise you, and a plan that can change underneath you.

CrescoDB runs on hardware you already rent. Your database is a standard Postgres, MySQL or SQLite file on a machine you control. Your app is a container on your own server. Nobody can pause it, and the bill is whatever you already pay your host.

The part most people have not seen

Because everything is a file on your machine, a coding agent can actually operate it. It can read your real schema, change it, call the API it just generated, and run the access audit to tell you whether any of it is unsafe to ship.

Against a hosted backend an agent is guessing, because the database is remote and the schema is behind a dashboard. Here it is not. That is an architectural difference, not a claim about whose model is smarter.

How it compares

Side by side against the managed platforms, written to be fair rather than flattering: CrescoDB vs Supabase, vs Firebase, vs PocketBase, vs Appwrite and vs Nhost.

What it costs

Local development is free forever. No account, no card, no trial. You pay when an app holds something you would hate to lose, and what you pay for is the operations half: backups we restore to prove they work, alerts when your app stops answering, and rollback points checked before you need them.