GitHub · Triage and routing · Tools and apps
realZachi/pg-jev
Seeing a decision model used as a database primitive rather than an API call.
What it does
A PostgreSQL extension that answers plain-language questions about your tables: Jev does the choosing, the database does the work. It puts the decision model inside the database rather than beside it, which is the same architectural bet Cloudflare made by putting Jev behind env.AI.run.
The repository's own description: “Ask your Postgres tables questions in plain language. A PostgreSQL extension powered by TypeSafe's Jev.”
How it works
It is a Postgres extension rather than an application — jev.control to register it, a Makefile to build it, sql/ for the extension body, test/ for its checks, a Dockerfile so it can be stood up locally, and META.json for PGXN-style packaging.
The repository, by the numbers
Read from the GitHub API on 2026-10-09. Stars and the last push move daily — quote them with the date, the way we do.
What we checked
- the README's quickstart and its measured numbers: CREATE EXTENSION jev, the jev()/jev_prob()/jev_choice()/jev_score() functions, 20 rows per request, and a 2,000-row table at ~3.5s / ~296k tokens / ~$0.012 on a first run
- META.json on master: jev 0.2.1, published to PGXN, requires PostgreSQL 14.0.0 and plpython3u, and declares the PostgreSQL licence
- the repository layout: jev.control, Makefile, sql/, test/, Dockerfile, META.json, CHANGELOG.md
- GitHub metadata on 2026-10-09: 1,069 stars, Shell, last push 2026-10-06, homepage pgjev.com
What we did not check. We have not installed the extension, wired it to a Jev key, or run a query through it, so every number here is the author's own measurement.