The Hasura GraphQL Engine exposes data from multiple sources through a single, composable, secured API endpoint, so an application — or an LLM app that needs grounded data — queries one place instead of each system separately.
The repository is a mono-repo carrying two engine generations. The v3 engine, in the v3 folder, powers Hasura DDN and supports PostgreSQL and its flavors, MongoDB, ClickHouse, and MS SQL Server. The v2 engine, in the v2 folder, is the current stable line.
DDN reaches data sources through Data Connectors, a separately published open-source set catalogued in a public Connector Hub. Custom business logic is written against Connector SDKs for TypeScript, Python, and Go, so logic that does not belong in the database still sits behind the same endpoint.
Features
- Single API endpoint: data from several sources served through one composable, secured endpoint
- Data sources: PostgreSQL and its flavors, MongoDB, ClickHouse, and MS SQL Server
- Data Connectors: source connectivity through open-source connectors listed in a public Connector Hub
- Connector SDKs: custom business logic written in TypeScript, Python, or Go
- Two engine generations: the v3 engine behind DDN and the stable v2 engine, kept side by side in one repository
