// the find
quarylabs/quary
Open-source BI for engineers
Quary is a dbt-style data transformation tool that lives inside VSCode, targeting engineers who want to manage SQL models, tests, and charts as code. It compiles Rust core logic to WASM so the extension works without a separate server process. Covers BigQuery, Snowflake, Postgres, Redshift, DuckDB, and SQLite.
The Rust-core-compiled-to-WASM approach is genuinely clever — the VSCode extension gets fast, dependency-light execution without shelling out to Python or a local server, which is dbt's Achilles heel for IDE integration. The protobuf-defined service contract between the WASM core and the TypeScript extension is a real architecture win: the boundary is typed and evolvable instead of string-serialized guesses. DuckDB support means you can run the full pipeline locally against flat files with zero infrastructure, which makes onboarding and local testing much faster than dbt. The DAG visualization built into the extension is the right place for it — seeing lineage without leaving your editor is genuinely useful.
2373 stars and only 56 forks after a YC batch is a weak adoption signal — dbt has 10x the ecosystem and Quary hasn't meaningfully dented it. Dashboards and Reports are still marked WIP in the README with no timeline, which means the 'BI for engineers' positioning is currently half-delivered. The VSCode-only interface is a hard constraint: if your team uses JetBrains or Neovim, you're back to the CLI only, with no web UI fallback. The last push was April 2026 but the star velocity suggests momentum has stalled — the repo doesn't show signs of an active contributor community beyond the founding team.