PrimDB Documentation
One platform for Deploy, Data, Auth, and AI, agent-native by design. Start here, then jump to the quickstart or a pillar.
PrimDB replaces the 6-8 separate services you normally stitch together to ship a project. Deploy, managed data, auth, and an agent API live behind one dashboard, one API, and one bill. It is fully managed on Hetzner in the EU. You get a primdb.dev subdomain by default, or connect your own domain.
Honest note. PrimDB is new and in early access. These docs describe what works today. Where a capability is planned or limited, the page says so plainly. No black box, no surprises.
The four pillars
- Deploy — push a repo, get a live URL. Preview per pull request, approval-gated promote, one-click rollback. See Deploy.
- Data — managed Postgres, Redis, full-text search, and object storage, provisioned in seconds. See Data.
- Auth — sessions, social sign-in, and org membership for your deployed app. See Auth.
- AI — an MCP endpoint your agents already understand, driving deploy, data, and feedback. See MCP.
Two ways to drive PrimDB
By hand: the @primdb/cli and the dashboard at dash.primdb.com. By agent: point any MCP client at mcp.primdb.com with a project token and it can do the same things. The surface is the same platform either way.
Where to go next
- Quickstart — install the CLI, log in, and ship your first deploy.
- CLI reference — every command group, global flags, and exit codes.
- MCP — connect Claude Desktop, Cursor, or any agent to your project.
- Preview feedback — the approval-gated preview and stakeholder overlay.
Every page here is also available as raw Markdown: append .md to any /docs/... URL (for example /docs/quickstart.md). A machine-readable index of the whole site lives at /llms-full.txt.