Skip to main content
Your app’s database needs are auto-detected. Postgres is provisioned automatically and DATABASE_URL is injected as an environment variable — no manual setup, no dashboard clicking.

​
How it works

  1. Database usage in your project is detected (Prisma schema, SQLAlchemy models, Drizzle config, etc.)
  2. A per-app Postgres schema is provisioned on a shared Cloud SQL instance
  3. DATABASE_URL is injected as an environment variable on deploy
  4. Connection pooling is handled via PgBouncer
No provisioning commands needed. Deploy and your database is ready.

​
Check database info

floo services info postgres --app my-app

Service: postgres (managed)
  Status:   active
  Host:     floo-db-abc123.us-central1.cloudsql.com
  Port:     5432
  Database: floo_my-app-uuid

DATABASE_URL is injected as an environment variable.

​
Access connection string

floo env get DATABASE_URL --app my-app

​
Supported ORMs

A standard DATABASE_URL is injected. Works with any ORM:
LanguageORMUsage
Node.jsPrismadatasource db { url = env("DATABASE_URL") }
Node.jsDrizzleconst db = drizzle(process.env.DATABASE_URL)
PythonSQLAlchemycreate_engine(os.environ["DATABASE_URL"])
PythonDjangoDATABASES = {"default": dj_database_url.config()}
Gopgxpgx.Connect(ctx, os.Getenv("DATABASE_URL"))

​
Limits

  • Query timeout: 5–30 seconds (configurable)
  • Connection pool: managed by PgBouncer
  • Storage: scales with your plan

​
For agents

# Check database status
floo services info postgres --app my-app --json 2>/dev/null | jq '.data'

# Get connection string
floo env get DATABASE_URL --app my-app --json 2>/dev/null | jq -r '.data.value'
Agents can provision a full-stack app with a database in one flow: 
floo deploy --json                                                        # Deploy app (DB auto-provisioned)
floo services info postgres --app my-app --json 2>/dev/null | jq '.data'  # Verify DB is ready
floo logs --app my-app --json --tail 5                                    # Check for connection errors