awesome-jev (Li-Evan)

How it uses Jev

Jev is TypeSafe's System One model. It answers typed questions about text with calibrated probabilities instead of generating prose, so code can branch, sort, and route on its judgments.

Use cases

• Jev experiments - Twenty-one low-latency demo apps, such as reranking 50 candidates in a single request.
• TypeSafe skills - Source of the agent skill, installable as a Claude Code plugin or through skills.sh.
• Building with Jev skill - Compact agent skill that teaches coding agents to structure programs around Jev.
• Jevify - Prompt to paste into a coding agent that has it study the Jev docs and your codebase, then find where cheap semantic judgments could cut cost and latency or enable new fe…
• Hello Jev (Real Python) - Companion code for Real Python's Get Started With Jev in Python video: a train-station help desk that swaps strict Y/N input parsing for a single Noul qu…
• AI cookbook for Jev - Nine runnable examples plus the four official patterns, written against the current Python SDK.

Getting started

- Jev Cheatsheet - One-page field guide to primitives, question design, confidence handling, limits, and tested SDK snippets. - Quick start - First request through the Playground, cURL, the Python SDK, or a coding agent. - Playground - Try a state and a set of questions in the browser before writing code (sign-in requ…