A small script runs the standard introspection query, normalizes the type list, and emits one row per type with slug, name, kind, description and a fields array. The fields array preserves arguments, type references and non-null wrappers. The resulting array is written to data/graphql-types.json for SleekRank to read on render.

Yes. SleekRank parses the JSON once and caches a slug-to-row map. Individual page render is an O(1) lookup, and WordPress full-page cache absorbs repeat traffic. A schema with ten thousand types is well within range, with no meaningful per-page slowdown.

Each field row carries an args array of argument objects with name, type and default value. The template renders them in a parameter table with the same conventions as GraphQL SDL, including non-null exclamation marks and list brackets. Because the data is structured, the rendering stays consistent across thousands of types.

Re-run introspection on deploy, overwrite the JSON, then sync SleekRank. Each type page picks up the new fields and arguments on the next request. Deprecated fields can be tagged with a deprecation reason that the template renders as a banner above the field row.

Yes. The /graphql-type/ index reads the array and groups by kind, with optional filters by owner_service. Both the index and the detail pages render from the same rows, so type counts, kinds and descriptions never disagree between list and detail views.

Union rows carry a memberTypes array with slugs of each constituent type. Interface rows carry an implementedBy array. The template renders both as link lists pointing at /graphql-type/{slug}/, so the type relationship graph is fully crawlable instead of being hidden behind JavaScript.

Yes. Many teams keep the explorer for interactive queries and add SleekRank pages for the static, indexable reference. Both can read from the same introspection JSON or the same upstream schema, so they stay aligned without duplicate maintenance.

InputObject types render the same way as Object types with a different kind badge. Query, Mutation and Subscription roots become first-class rows too, with their top-level fields and arguments rendered exactly as the schema declares them.