Orbital

Orbital automates the integration between your data sources. There is no central mapping layer to write. It reads the API specs your teams already publish and builds each integration at query time, which removes the stitching code that breaks whenever one of those APIs moves.

The metadata comes from Taxi, a language that embeds semantic types in OpenAPI, Protobuf, Avro and JSON Schema documents, and Orbital itself is a TaxiQL query server. Specs reach it by polling, from a Git repository, or pushed by a service on startup. A query describes the response you want and Orbital composes the calls that produce it.

The project gives several reasons for working this way.

• No glue code: integration code between APIs is brittle and breaks every time one of them changes.
• Spec first: existing API specs drive the integration, leaving less code to maintain.
• Technology agnostic: REST, gRPC, SOAP, Kafka and databases are all treated the same way.
• Automatic adaptation: as specs change, integration flows adjust and consumers are left alone.
• Resolver free: unlike GraphQL federation, there are no resolvers for a middleware team to maintain.
• Decentralized: teams evolve independently, without sharing identifier schemes or common models.

Getting started is a docker compose file fetched from the project's start page, then a browser pointed at the local port it exposes. Orbital fits an organization that wants federation across mixed technology without moving its stack to GraphQL.