firebase-data-connect

Run npx -y firebase-tools@latest init dataconnect to scaffold the dataconnect directory with dataconnect.yaml, schema.gql, and connector files. Then define your schema, write operations, and deploy with firebase deploy --only dataconnect.

Use the @auth directive with levels PUBLIC, USER, or NO_ACCESS on each operation. Add @check and @redact for row-level security and validation, and reference auth.uid in expressions to scope data to the current user.

Default to Native GraphQL for standard CRUD, filtering, and joins since it provides type safety. Use Native SQL only for PostgreSQL extensions like PostGIS, window functions, complex aggregations, or data-modifying CTEs.

Yes, single-entity lookups refresh automatically, while list queries use the @refresh directive for event-driven updates via onMutationExecuted or time-based polling with every. Clients consume updates through the subscribe() method in generated SDKs.

Configure the SDK output in connector.yaml with javascriptSdk, kotlinSdk, swiftSdk, or dartSdk blocks, then run firebase dataconnect:sdk:generate. Regenerate and redeploy whenever operations change, since operations are stored server-side.

The onMutationExecuted trigger fires again if your function executes another mutation. Always filter triggers by service and operation, and ensure no mutation inside the handler can retrigger itself directly or indirectly.