gqless

Yes! We announced experimental support earlier and here's the new spec we're implementing that will support all databases and remote schemas too. https://github.com/hasura/graphql-engine/issues/6991 General support for inherited roles is one of the things I'm most excited about because it makes a bunch of hard things around reusing and composition so easy. This improvement plays really well along with things like "role-based schemas" so that GraphQL clients have access to just the exact GraphQL schema they should be able to access - which is in turn composed by putting multiple scopes together into one role. Also interesting is how well this could play with other innovations on the GraphQL client ecosystem like gqless[1] and graphql-zeus[2] because now there's a typesafe and secure SDK for really smooth developer experience on the client

On the GraphQL side you can use gqless[0] (or the improved fork I helped sponsor, here[1]). It's by far the best DX I've had for any data fetching library: fully typed calls, no strings at all, no duplication of code or weird importing, no compiler, and it resolves the entire tree and creates a single fetch call. [0] https://github.com/gqless/gqless [1] https://github.com/PabloSzx/new_gqless

> Magic is great until you run into an edge case, or try to do anything complicated Yep totally agree. While it's not complete yet, (see https://github.com/samdenty/gqless/blob/master/README.md#gql...) I've designed with the sole focus for being used inside complex applications. The "magic" part is only a small portion of gqless. For 1.0, there will be an option to "use plain queries". I've tried re-inventing everything - but it's also innovating in the reactive, query-optimization, local state and cache areas. It can do even more optimizations than Relay ;)