dperezcabrera/system-one-chess
User repository on GitHub
Chess against Jev, TypeSafe AI's System One model, through OpenRouter. Built with the pico framework.
- Forks
- 1
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Demos & Experiments
System One Chess
Browser chess lab where Jev picks from legal moves as one choice question per turn, alongside LLMs and local open models, with Swiss tournaments, live cost tracking, and in-browser Stockfish analysis.
- Stars
- 4
- Language
- Python
- License
- GPL-3.0
- Published by
- Community
- Author
- dperezcabrera
- Last verified
- Oct 2026
About this project
ai-chess-lab: LLMs, System One models and you play chess in the browser, one game at a time or in Swiss tournaments and full leagues, with every move, token, second and dollar on the record and Stockfish analysing the games in your browser. The first battle, 16 players and 72 games, is in the repository.
Quoted from the project's README. Read the full README.
Used inJev for Software & DevTools
- Languages
- Python
Highlights
- Each turn is one choice question whose options are exactly the legal moves.
- Swiss rounds or full leagues with standings, per-move records, and PGN export.
- Live footer tracks calls, tokens, latency, illegal answers, and dollars per game.
- Stockfish in the browser scores games locally with evaluation charts and blunder counts.
Watch out
Jev is a fast classifier here, not a chess engine, so expect plausible moves rather than strong play. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
A position has at most a couple hundred legal moves and a choice supports up to 255 options, so one request always fits. The model can only select from the supplied legal moves.
Run the published container image and open the tournament review page: the bundled multi-game tournament replays with no key. Keys are only needed to play new games or run your own tournaments.
The footer shows calls, tokens, latency, illegal answers, and dollars from gateway usage. The documented rate is a fractions-of-a-cent cost per move with sub-second latency per Jev turn.
Yes. One open model runs locally with no per-move cost after the download, and another family answers through the same decision protocol via a demo endpoint or self-hosted server with slower shared-GPU timing.
Run docker run --rm -p 127.0.0.1:8000:8000 ghcr.io/dperezcabrera/ai-chess-lab:latest, then open http://localhost:8000/tournament; the bundled 72-game September 2026 tournament replays with no key, and new games need a gateway key.
Tournaments save to /app/tournaments and vanish with the container, so mount a folder there; copying results/20260922-141451-7ccf.json in keeps the bundled tournament too.
The footer reads calls, tokens, average latency, illegal answers, and dollars from the gateway's usage data, one row per model with a total; Jev on both sides is about $0.00005 per move at 300-400 ms.
More Jev resources like this
4GitHub stars
Browser-use layer for CLI coding agents: Jev picks the operation and target element in one round trip per step, exposed over MCP with a CLI, a Python API and an HTTP server.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
brnyxx/jev-raPython
brnyxx/jev-raPython1GitHub stars
An open Apache-2.0 2B non-autoregressive decision model for agent action routing on a Qwen3.5-2B backbone: one forward pass over a candidate set with a pointer head and an escalation head, shipped as BF16, FP8 and INT8 weights.
Jev Demos & Experiments#community#awesomejev#catalog-import
Community
mkeco/Cerebellum-2BPython
mkeco/Cerebellum-2BPython1GitHub stars
Browser agent that takes one Jev decision per action on top of agent-browser and only calls an LLM when confidence drops, escalating instead of guessing, with values, origins and irreversible clicks all guarded in code.
Jev Demos & Experiments#community#awesomejev#catalog-import
Community
HAR5HA-7663/hunchPython
HAR5HA-7663/hunchPython0GitHub stars
A local Python browser automation UI that turns a URL and a plain-language goal into bounded actions, with Jev picking the next operation from observed candidates and a separate text model used only to fill fields.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
KesavanKing/jev-browserPython
KesavanKing/jev-browserPython1GitHub stars
A Streamlit Tetris where the engine enumerates every legal placement, Jev picks one typed placement ID, and Python verifies rotation, column and exact occupied cells before the piece is allowed to lock.
Jev Demos & Experiments#community#jevbest#catalog-import
46GitHub stars
Open 150M encoder that reproduces the typed noul, choice, and score interface in one forward pass, trainable on a free GPU and runnable locally for nothing.
Jev Demos & Experiments#community#jevbest#catalog-import
From the Jev guides
Original write-ups on what this entry is for.
A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.
Jev Demos & Experiments#gaming#agents#control
Read article
From the Jev community
Posts from builders shipping with Jev right now.