Developer docs
The managed backend explained end to end: data, auth, functions, connectors, deployment and hosting.
ExploreDatabase, auth, realtime, integrations and hosting ship with every project. Deploy from your IDE or hand the whole thing to your AI coding agent — the CLI and typed SDK speak both languages fluently.
Start building Read the docsDescribe the app in natural language; the CLI scaffolds the project, generates the data model, wires auth and deploys.
Nine building blocks come with every project. None of them needs a separate provider, a separate bill or a week of setup.
The CLI is optimized for coding agents like the one you already use: skills and context files cut the rework, and a session auth check runs before anything touches your project.
Google sign-in ready on day one, email and password built in, custom auth flows when you need them. No separate identity provider to wire up.
Your data model lives in code as schema files: typed, reviewable, versioned. Storage, validation, migrations and queries are handled for you, with full TypeScript support.
AI, email, files and images built in, plus one-click OAuth connectors for the services your app talks to. See the integrations catalog.
Secure server-side code for the logic that must not run in a browser — with secrets kept server-side and external APIs called safely.
Row-level security is enforced across every client — web, mobile or API — not merely hidden in the interface.
Data changes reach every connected client live. No WebSocket configuration, no polling loops to babysit.
CDN-backed uploads with permission-aware access: public assets and private files in the same project.
One command deploys frontend, functions and schema together — CDN hosting, custom domains and automatic HTTPS included.
The backend stands on its own. Attach the frontend you want, or none at all.
Data, auth and billing-ready backends for products you charge for.
Auth, data and AI for your extension — no server of your own to run.
Services and automations that other systems call, with no UI at all.
Bots for the chat platforms your users live in, running on your backend.
Dashboards and operations tools your team uses every day.
How an AI coding agent ships a backend here — the same flow you can run yourself from any terminal.
npx gatebuild create my-app scaffolds a standard frontend plus a config folder: project settings, entity schema files and function entry points.
The agent writes entities as plain schema files — readable, diffable and reviewable like any other code in your repo.
npx gatebuild deploy ships schema, functions and frontend together and returns a live URL on your project’s subdomain.
Ask for file attachments and an admin role; the agent updates the schema and permissions and redeploys. The conversation is the commit history’s first draft.
The platform runs your backend so you do not have to. The code and the data stay yours, and the exit door is always unlocked.
Push from the online editor, pull in your IDE. Merges on your main branch sync back into the project; publishing stays a separate, deliberate action.
Builder+Download the complete project — frontend and backend — whenever you want a copy that answers to nobody’s uptime.
Run the project locally with the same SDK. Your data model is plain schema files, not a proprietary format, so leaving never means rewriting.
The editor is one client of the platform. Your scripts and your coding agents are first-class clients too.
External AI assistants can create and change apps through our MCP server, working under the same permissions as a human editor.
A programmatic mirror of the editor: data model, branding, secrets, GitHub, SEO, analytics, payments, the marketing agent, social content and ads.
Hand a project to your own external coding agent through the CLI or MCP. Its reads, edits and commands are metered in integration credits; the agent’s own AI runs on your subscription, and no message credits are involved.
Three references cover the whole developer surface.
The managed backend explained end to end: data, auth, functions, connectors, deployment and hosting.
ExploreCreate, sync entities and code, and deploy everything with one command — including apps that started in the online editor.
ExploreData, auth, agents and integrations from any frontend or backend — the same typed SDK that powers AI-generated apps.
ExploreA complete managed backend — database, auth, realtime sync, file storage, backend functions, integrations and hosting — defined in code and configuration files in your project, and driven end to end through one CLI.
Any frontend framework, or no frontend at all. Server-side code is TypeScript-class functions on a managed runtime, and the typed SDK and CLI work from any environment your agent or IDE runs in.
Yes. Run functions, entities and uploads locally, or connect a local frontend to the hosted backend while you iterate — then deploy when it is ready.
They drive the CLI directly. A skills package gives the agent the platform context it needs up front, so it scaffolds, edits schemas and functions, and deploys with far less trial and error.
Yes. Export or eject the project locally, sync two ways with GitHub, and call the same API and SDK the generated apps use. Managed hosting is a convenience, not a lock.
No. Backend-only projects — entities and functions with any client connected through the JS SDK — are a first-class template, made for headless APIs and automations.
Free to start, with usage metered in credits from one wallet shared across building and running. See the pricing page for the tiers.
The backend is the same one AI-generated apps run on: row-level security enforced for every client, managed deployment with automatic HTTPS, and your choice of custom domain.
One command scaffolds it. One command deploys it. Your agent can do both.
Start building