Published event
DeveloperTools PolicyChange 1 source(s)

Jevstiller: Distill a repeated Jev classification task into a local model

Updated October 1, 2026 · 12:00 AM · 21 · source date September 30, 2026

Summary

Jevstiller: Distill a repeated Jev classification task into a local model tomerglick57 / Jevstiller Public Notifications You must be signed in to change notification settings Fork 4 Star 46 Branches Tags Open more actions menu Latest commit History 71 Commits 71 Commits Folders and files Name Name Last commit message Last commit date .github .github benchmarks benchmarks brand brand deploy deploy docs docs examples examples experiments experiments jevstiller jevstiller tests tests website website .dockerignore .dockerignore .env.example .env.example .gitignore .gitignore CHANGELOG.md CHANGELOG.md CODE_OF_CONDUCT.md CODE_OF_CONDUCT.md CONTRIBUTING.md CONTRIBUTING.md DEPLOYMENT_PLAN.md DEPLOYMENT_PLAN.md DESIGN.md DESIGN.md Dockerfile Dockerfile LICENSE LICENSE NOTICE NOTICE README.md README.md SECURITY.md SECURITY.md build-requirements.in build-requirements.in build-requirements.txt build-requirements.txt docker-compose.yml docker-compose.yml pyproject.toml pyproject.toml uv.lock uv.lock Repository files navigation Jev, distilled on the fly. Put Jevstiller in front of a repeated Jev classification call.

Why it matters

This PolicyChange is relevant to the technology intelligence record because it involves GitHub, OpenAI, Amazon, Docker. The source article should remain the factual reference for follow-up coverage.

Key facts
  • tomerglick57 / Jevstiller Public Notifications You must be signed in to change notification settings Fork 4 Star 46 Branches Tags Open more actions menu Latest commit History 71 Commits 71 Commits Folders and files Name Name Last commit message Last commit date .github .github benchmarks benchmarks brand brand deploy deploy docs docs examples examples experiments experiments jevstiller jevstiller tests tests website website .dockerignore .dockerignore .env.example .env.example .gitignore .gitignore CHANGELOG.md CHANGELOG.md CODE_OF_CONDUCT.md CODE_OF_CONDUCT.md CONTRIBUTING.md CONTRIBUTING.md DEPLOYMENT_PLAN.md DEPLOYMENT_PLAN.md DESIGN.md DESIGN.md Dockerfile Dockerfile LICENSE LICENSE NOTICE NOTICE README.md README.md SECURITY.md SECURITY.md build-requirements.in build-requirements.in build-requirements.txt build-requirements.txt docker-compose.yml docker-compose.yml pyproject.toml pyproject.toml uv.lock uv.lock Repository files navigation Jev, distilled on the fly.
  • Put Jevstiller in front of a repeated Jev classification call.
  • At first every request still goes to Jev.
  • From Jev's own answers — with their full probability distributions — it trains a small local model on your traffic, checks that the model agrees with Jev within a budget you set, and then answers most requests itself.
  • Uncertain or novel input, and a permanent random audit slice, keep going to Jev.
  • Run Jevstiller next to your service and point the SDK at it; nothing else changes: docker run -d -p 8080:8080 -v jevstiller-data:/data ghcr.io/tomerglick57/jevstiller export TYPESAFE_BASE_URL=http://localhost:8080 # your service keeps its own TYPESAFE_API_KEY The recording above is examples/proxy_demo.py : 5,000 real customer messages through the unmodified SDK, 8 threads, against live Jev.
Entities in this story
Related events