Playground
CommunityJev Demos & Experiments
Jev Decision Wall
Browser demo that sends one text to TypeSafe Jev and lights up 200 yes/no judgments as a live heatmap in roughly 150-300 ms, with seven 200-question sets, a threshold slider, and zero npm dependencies.
- Stars
- 0
- Format
- Playground
- Reading time
- 2 min
- Published by
- Community
- Added
- Nov 2026
Highlights
- 200 yes/no questions answered in one parallel call to TypeSafe Jev in about 150-300 ms
- Seven question sets of 200 unique questions each; all seven return 1,400 answers in 283 ms
- Zero npm dependencies, Node 20+, MIT licensed, served on port 8770 or via Docker
- The server holds the TypeSafe key and the browser never calls TypeSafe directly
- A full 200-question wall costs roughly $0.0002 at $0.042 per million input tokens
Questions about this entry
Short answers drawn from the linked project or page.
Yes. The server reads TYPESAFE_API_KEY from the environment or a .env file, or from a key file set by JEV_KEY_FILE, and adds it to each request it makes to TypeSafe Jev.
Jev charges for input tokens only. A 200-question call is about 3,000-4,700 tokens, so at TypeSafe's published $0.042 per million input tokens a full wall costs roughly $0.0002. The server stops calling at a daily cap, default 5M tokens or about $0.21.
The browser never talks to TypeSafe directly, because the API does not allow cross-origin calls. The Node server adds your key, never sends it back to the page, and logs every call to ~/.config/jev/usage.jsonl.
There are seven sets, each exactly 200 unique questions arranged in 8 categories of 25, covering everyday messages, scams, farm notes, research computing, ops and security, news claims, and story pitches. Firing all seven is 1,400 questions across seven parallel calls, measured at 283 ms wall clock.
The project has zero npm dependencies and requires Node 20 or newer. It is released under the MIT license.
Yes. Sets are plain JavaScript files in public/sets/, with each line formatted as a short label followed by a full yes/no question. The README estimates about an hour of work, and npm run check validates the question count and uniqueness.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Text-to-screenshot (Malay Vasa)
X: Google's MediaPipe Decision Maker, tested in the browser
X: Playwright CLI + Jev vs Playwright MCP
X: jev-ultrafast splits every browser decision into operation, target and value. JEV passed 18,000 stars in its…
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.