jev-plays-pokemon-red

Pokemon Red on PyBoy where code owns the route and the arithmetic, Jev only picks at branches, and every battle turn's faint prediction is scored by Brier against RAM state.

Pokemon Red on PyBoy: code owns the route and the arithmetic, Jev picks at branches in about 100 ms, calibration measured instead of assumed

How it uses Jev

The catch, up front: it cannot plan. It never sees more than the current decision. Every route, threshold and piece of arithmetic is ordinary Python, and the model only picks among actions the code already proved legal. It is also not asked often. Code handles the routine ticks and only calls Jev where the game actually branches. Between the bedroom and Oak's lab it is not asked anything at all.

Source: community catalog. Details and metrics reflect the published community snapshot and are not a performance endorsement. Documentation details extracted from GitHub · the README.

Keep exploring

Similar to jev-plays-pokemon-red

Browse all
jev-snake project preview
Games & simulations
An experimental Snake environment where the game engine owns deterministic rules and TypeSafe AI's Jev makes the movement decision from structured state on every tick.
Jev plays Mario project preview
Games & simulations
Less talking. More jumping. Watch Jev navigate Super Mario Bros. from game state.
mario-jev project preview
Games & simulations
Python prototype that plays NES Super Mario Bros. from structured RAM observations, with Jev answering focused movement and jump questions.
jev-aita project preview
Benchmarks & research
Benchmark of TypeSafe's Jev against Sonnet 5, GPT-5 nano and local LLMs on 770 Reddit AITA verdicts: Brier scores, latency and cost
Browser & computer use
A browser agent that turns a single goal into a sequence of lightning-fast actions.
openJev-verdict-2.0 project preview
Jev alternatives & open source
Calibrated 151M Non-Autoregressive Decision Engine beating TypeSafe Jev & Laya on LocalLLaMA/typed-decisions (77.10% acc, 0.0636 Brier, 0.0144 ECE)
prompt2jev project preview
Agent tooling
Agent skill and CLI that turn natural language, an LLM prompt, or the code that runs one into a TypeSafe Jev decision: typed state, Choice/Score/Noul questions, and a runnable script
jevmeter project preview
Applications
Put a live Jev (TypeSafe) meter on any video: every sentence scored, rendered as a 16:9 edit
simple-jev project preview
Applications
Turn any open model into a classifier/jev endpoint