nice, that's great to hear! The plan is to push towards being the best TypeScript semantic layer for ClickHouse apps. The big gaps I’m focused on are governance, contract snapshots/diffs, relationship-aware queries, richer metrics/time semantics, BI/GraphQL connectivity and better generated tooling.
The biggest hurdle is that it’s still a young library sitting in the core read path for production analytics. That naturally requires early adopters who believe in the direction, are willing to use in real systems, and can be patient whilst gaps are closed.
The base is solid though: the underlying query builder has 10k+ weekly downloads and production usage. The newer surface is the semantic layer itself.
ClickHouse has incredible performance but the TypeScript tooling is years behind Postgres.
You end up with SQL strings scattered across services, types that drift when your schema changes, and the same metric logic duplicated everywhere.
hypequery generates TypeScript types directly from your ClickHouse schema and lets you define queries once, then run them over HTTP, in React hooks, in background jobs, or from AI agents.
Would love feedback from anyone actually running ClickHouse WITH TypeScript in production.