Principles
- GraphQL semantics come first. Python ergonomics must compile to behavior that remains legible in GraphQL terms.
- Simple definitions should stay simple. Annotations are the default path; metadata is available without forcing boilerplate.
- One concept should have one contract. Type-like decorators share definition
behavior, and
Fieldworks consistently across object and operation classes. - The runtime should be inspectable. Parsing, validation, schema compilation, execution, SDL, and introspection remain explicit stages.
- Project claims must match implemented capability. Experimental and unsupported features are documented as such.
Scope
FastQL owns a Python schema definition API and an in-process GraphQL language and execution engine. It includes local tooling needed to inspect and exercise that model. Production HTTP framework integration, streaming subscription transport, federation, distributed tracing backends, and deployment infrastructure are not currently core capabilities. Those features should enter through explicit specifications rather than implicit compatibility claims. The canonical requirements live underopenspec/specs.