Quickstart
Create a project, copy the connection string, and ship without learning a custom platform first.
1. Choose an organization
Sign in, then select or create an organization from the dashboard header. Projects, billing, API keys, and webhooks all belong to the organization.
A subscription (Vibe, Ship, or Business) is required before provisioning - plans are billed per organization and project entitlements are derived from the org's plan. See Limits for what each plan includes.
2. Create a project
Open the Projects view, create a project, and choose the target region - it is fixed after creation. The control plane enqueues a provision job; the project flips to ready when the database, role, and pooled access exist.
3. Connect
When the project is ready, copy either the direct URL or the pooled URL from the connection details and drop it into your environment variables.
psql "$DATABASE_URL"Rule of thumb: pooled URL for app traffic, direct URL for migrations and admin sessions. Details in Connections.
4. Optional CLI flow
If you want the CLI to write local env vars for you:
brew install capy-base/tap/capydb
capydb login
capydb link --project my-appThe CLI opens a browser login flow, uses the active dashboard organization, detects your framework (Next.js, Prisma, Django, Rails, ...), writes DATABASE_URL / DATABASE_DIRECT_URL / DATABASE_POOL_URL into the right env file, and gitignores the file holding credentials. The full command set is in the CLI reference.
5. Iterate without panic
- Create disposable preview databases for test and review flows
- Create backups before risky changes and restore from backup keys, timestamps, or pinned restore points
- Import an existing database from a source URL when moving from another host - run the preflight first
- Rotate project credentials without rebuilding the project
- Wire Vercel, Netlify, or the GitHub Action so previews and env vars manage themselves
Documentation
CapyDB is managed Postgres built on one primitive - the isolated database cell. Standard direct and pooled connections, disposable previews, backups and PITR restores, imports, integrations, webhooks, and a real API. No proprietary SDK required.
Set up with an AI assistant
Paste one prompt into any AI assistant and let it provision a CapyDB Postgres database - you approve access once in a browser.