Organization repository on GitHub
Jev + Arrow
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Repos & SDKs
Jevaro
A Python batching proxy for Jev that returns Apache Arrow streams, with Python and JavaScript SDKs. It packs many independent states and questions into each TypeSafe API call or runs them concurrently, then emits the answers as an ordered Arrow IPC stream with the schema first.
- Stars
- 16
- Language
- Python
- License
- APACHE-2.0
- Reading time
- 2 min
- Published by
- Community
- Author
- columnar-tech
- Last verified
- Oct 2026
Used inJev for Data & Analytics
- Languages
- PythonJavaScript
- Platforms
- API
- Capabilities
- Open sourceSelf-hostedRequires API key
Highlights
- Ask the same questions about many independent states in one client request, run concurrently or packed with rows_per_call into fewer calls.
- Each state becomes a row and each question a column, in input order, and the schema arrives before the answers.
- Arrow extension types preserve the meaning of Choice, Noul, Score and Box, with shared labels and legends stored once in the schema.
- Three packages: jevaro-server (the proxy), jevaro for Python with sync and async readers, and jevaro for JavaScript in browsers and Node.js.
- Quickstart needs Python 3.11+ and a TypeSafe API key; results go to a .arrows file readable by PyArrow, pandas, Polars or DuckDB.
Quickstart
bash
python3 -m venv .venv
.venv/bin/python -m pip install jevaro-server jevaro
export TYPESAFE_API_KEY="your-api-key"
.venv/bin/jevaro-serverWatch out
README reviewed 2026-10-17; the digest truncates mid-sentence after the documentation list, so server configuration beyond rows_per_call, error handling and pricing were not captured. The project is described as an experiment in bulk inference, and every evaluation uses the TypeSafe API and incurs its normal charges.
Questions about this entry
Short answers drawn from the linked project or page.
A Python batching proxy for Jev by columnar-tech that returns Apache Arrow streams, with small Python and JavaScript SDKs for reading them; the README calls it an experiment in bulk inference.
The TypeSafe API evaluates one state per request, so Jevaro makes those calls concurrently or packs many states into each call with rows_per_call, letting you ask the same questions about many independent states at once.
Each state becomes a row, each question a column, in input order. Extension types carry Choice, Noul, Score and Box meanings, with shared labels and legends stored once in the file's schema metadata.
With Python 3.11+, create a venv, install jevaro-server and jevaro from PyPI, export TYPESAFE_API_KEY, start the server, then run the example script that evaluates three customer messages.
Any Arrow consumer: the README names pandas, Polars and DuckDB, and the reader script for the saved .arrows file needs only PyArrow to print the full answers using the schema metadata.
The README does not state proxy pricing in the captured text; every evaluation uses the TypeSafe API and incurs its normal charges, as the README notes next to the example run.
Yes; jevaro for JavaScript provides async Arrow readers for browsers and Node.js, and the quickstart shows how to inspect the schema and run the same example from JavaScript.
More Jev resources like this
A System One layout optimizer for decision-model workloads: it reorders rows and fields so a prefix-caching backend reuses input computation, then returns per-record decisions in original order. Its bundled backend serves AutoTrust JEV-9B through vLLM.
Jev Repos & SDKs#community#github#python
Community
0814wdwd/solo_jevPython
0814wdwd/solo_jevPythonA Python library shaped like the re module that finds, replaces, splits and extracts text by meaning: loose finders propose spans and Jev judges each one with a probability. Every match stays a verbatim slice of the input with its offsets, so it can be wrong but never invented.
Jev Repos & SDKs#community#github#python
Community
nothans/ex-regexPython
nothans/ex-regexPythonA Python library that turns Jev-style decision models into ordinary function calls: check returns a bool, classify picks one option, label picks every option that applies, and score rates on a scale, with probabilities behind every answer.
Jev Repos & SDKs#community#github#python
Community
jlowin/vibecheckPython
jlowin/vibecheckPythonHTTP server for the Mica v0.5 4B Preview decision model: GGUF weights run through llama.cpp and answer noul, choice and score questions with calibrated probabilities over the TypeSafe /v1/systemone format.
Jev Repos & SDKs#community#github#python
Community
akivet/MicaPython
akivet/MicaPythonA search API over about 203 million OpenAlex English title and abstract records, running DenseOn embedding search by meaning and a tantivy index for exact words behind one HTTP service. Its whole-field topic maps use TypeSafe's Jev to judge archetype and cited papers.
Jev Tools & CLIs#community#github#python
Community
maximerivest/srl-searchPython
maximerivest/srl-searchPythonServing code for the Jeff 1.0 family of Jev-style decision models: a state, a typed question and its options go in, and one forward pass returns a probability for each option. Four open models from 4B to 31B parameters are served the same way.
Jev Repos & SDKs#community#github#python
Community
j-geuter/jeff-servePython
j-geuter/jeff-servePythonFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.