How does it work?
gqless creates an object using information from your schema. Accessing new properties adds them to the current "commit". GraphQL queries are then generated every few milliseconds.
The response from the query is merged into the cache. Now when you access the same properties again, the newly fetched values are returned.
How will it affect bundle size?
In the code-generated files, an object representation of the schema is outputted - negligible for most apps. This could be optimized in the future by stripping out unused parts.
The gqless packages are lightweight and don't require external dependencies.
How fast is it?
Very fast, gqless's architecture was designed with performance in mind. Unlike other clients, queries are generated directly - instead of outputting intermediate AST.