What it is
An Open WebUI Tool that lets a chat model send typed questions to TypeSafe's Jev and receive calibrated probabilities instead of guesses. Aimed at users wanting defensible scores, fast structured decisions, batch labelling, and rubric grading.
How it uses Jev
The model calls tools like jev_yes_no, jev_classify, jev_score, ask_jev, or jev_classify_batch. State comes from an argument, attached files, or recent chat messages. Jev returns probabilities, distributions, and confidences, which the tool renders as summaries the model can reason over.
Primitives:choicescorenoul
Technique worth stealing
Bundle several typed questions about one state into a single Jev request, evaluated in parallel.
Try it
Paste ask_jev.py into Open WebUI Workspace Tools, set JEV_API_KEY, enable on a model.
Evidence
Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.
- Jev-centricyes0.94
- Shows a System One patternunclear0.56
- Handles uncertaintyno0.06
- Measuredno0.06
- Runnableyes0.86
- Worth recommendingunclear0.47
- Model replicano0.03
- Problem scopescore on a 0–2 scale1.04
- About Jevyes0.98
Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.