Jev in production › SQL and data pipelines
Jevaro
Jevaro batches Jev requests concurrently and returns typed answers as Apache Arrow IPC streams for pandas, Polars and DuckDB pipelines.
- Use
- SQL and data pipelines
- Industry
- Data and analytics
- Form
- Open-source tool
- Stage
- Beta
- Plugs into
- pandas, Polars, DuckDB
- Listed
- 2026-09-30
- Found via
- github
- Repository
- columnar-tech/jevaro
- Stars
- 16
- Forks
- 0
- Last push
- 2026-10-05
- Language
- Python
- License
- Apache-2.0
The README opens with
A Python batching proxy for Jev that returns Apache Arrow streams, with Python and JavaScript SDKs.
Jevaro is an experiment in bulk inference: ask the same questions about many independent states in one client request. The TypeSafe API evaluates one state per request. Jevaro makes those calls concurrently, or packs many states into each call with rowspercall. It converts the JSON answers into an Arrow IPC stream, in input order.
The schema arrives before the answers. Each state becomes a row and each question a column. Arrow extension types preserve the meaning of Choice, Noul, and Score, with shared labels and legends stored once in the schema. The resulting Arrow data can go into pandas, Polars, DuckDB, and other Arrow tools.
Badge
For the project's own README, linking back here:
Also used for sql and data pipelines
- JevSDSQL $0.389
- grev sub-cent cost per few thousand lines
- Jev IDS 7.7x
- jev-harness (AntonioCoppe) 1.3 seconds
- jev-ultralightspeed 41% less money
- jlink $2.95
- signal-weave 130/144 exact outcome decisions
- jevgrep mean F1 of 0.90