What it is
Reads up to 1,500 posts from the last year of a public Telegram channel and draws a shareable card showing the mix of post kinds, ads, clickbait and emotional pressure, and how it changed monthly.
How it uses Jev
Each post with text goes to Jev in its own request with four typed questions: the kind of post (one of ten), is it a paid ad, does it use clickbait, does it push on emotions. Answers are aggregated into shares, monthly mix, median views per kind and most telling posts.
Primitives:choicenoul
Technique worth stealing
Browser-driven run: the Worker plans pages, the browser fetches them in parallel, and each page is judged six posts at a time.
Try it
git clone, npm install, cp .env.example .env, paste a key, npm run dev, open http://localhost:5190.
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-centricyes0.98
- Shows a System One patternyes0.86
- Handles uncertaintyno0.05
- Measuredyes0.74
- Runnableyes0.93
- Worth recommendingunclear0.59
- Model replicano0.03
- Problem scopescore on a 0–2 scale0.57
- About Jevyes0.98
Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 21 Sept 2026.