User repository on GitHub
Demos to test the effectiveness of TypeSafe's "Jev" System One Model
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Demos & Experiments
Jev Demos
A two-package collection of TypeSafe System One demos: jev_common is a shim for the API that loads .env and records results, and maze_lookahead checks spatial reasoning by testing how far the model can think ahead in a maze.
- Stars
- 0
- Language
- Dart
- Published by
- Community
- Author
- Bud-ro
- Last verified
- Oct 2026
About this project
A collection of demos utilizing the system one models from TypeSafe AI. Put your API key into .env and then follow demo specific examples.
Quoted from the project's README. Read the full README.
Works forJev for Classification
Used inJev for Software & DevTools
Highlights
- jev_common is described as a shim for the TypeSafe API plus .env loading and result recording, shared by the demos.
- maze_lookahead is the only demo listed: checking spatial reasoning and its ability to think ahead.
- Setup is one step: put your API key into .env, then follow the demo specific examples.
Watch out
The README is a one-line description, a two-row package table and a .env instruction, and nothing more. There is no documented test command, no license section, no recorded results and no sample output, so the maze demo's prompts, scoring method and dependencies are not visible from the README, and jev_common is described only as a shim for the API with .env loading and result recording. The follow-along material lives inside the demo directories rather than in the README, so expect to read the code to judge whether the spatial reasoning check means anything. 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.
One demo and one helper. The table lists packages/jev_common, a shim for their API with .env loading and result recording, and packages/maze_lookahead, which checks spatial reasoning and its ability to think ahead. Nothing else is listed, and the README does not say whether more demos are planned or which model identifiers the collection targets beyond the system one models.
The README only says it checks spatial reasoning and the ability to think ahead. There is no metric, threshold, benchmark, seed or sample output documented, and jev_common's result recording has no documented format or file location, so what a good or bad run looks like is only discoverable by reading the demo source.
Neither. The README is a one-line description, a two-row package table and one setup step: put your API key into .env and then follow the demo specific examples. No runtime version, package manager, install command, lockfile, test or lint script is named, and there is no license section, so which language and runtime these packages target is only discoverable from the repository files themselves.
Neither is stated. The README covers only the project description, the package table and the .env setup step, with no license section, changelog or maintenance notes. Treat the terms as unclear until you check the repository files, and do not assume the demos follow any particular release cadence.
As a sketch at most. jev_common is described as a shim for their API with .env loading and result recording, so those two are the transferable parts. What a run sends, which questions it asks, what a recorded result looks like and where it is written are all undocumented, and these are demos rather than a maintained package surface.
Only decision models. The one-line description says a collection of demos utilizing the system one models from TypeSafe AI, and the package table holds exactly two entries. There is no mention of generation, agents, tools, evaluation or any other product, and nothing indicating the collection is meant to grow beyond what is listed.
The README does not say. It has a one-line description, a two-row table naming packages/jev_common and packages/maze_lookahead, and one setup instruction. Whether there are scripts, notebooks, tests, a license file or a workspace root above packages/ is not described anywhere, so the shape of the repository has to be read from the file tree itself.
More Jev resources like this
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
941GitHub stars
Library that adds schema-guaranteed typed outputs to existing autoregressive LLMs without changing weights, served through SGLang or a hosted API.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
TypeLLM/TypeLLMPython
TypeLLM/TypeLLMPython98GitHub stars
Self-hosted Docker service and web UI that packages Laya structured-decision inference behind a compatible System One endpoint.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
1Panel-dev/laya-serverPython
1Panel-dev/laya-serverPython39GitHub stars
Mac menu-bar voice controller where Jev picks the next action from observed Accessibility controls and Swift executes it in an observe-decide-execute loop.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
ronadin2002/jev-cuaSwift
ronadin2002/jev-cuaSwift55GitHub stars
Toolkit that makes a model behind an OpenAI-compatible endpoint or running in-process answer typed choice, score, and yes/no questions from next-token probabilities in Jev API shape.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
fidecastro/jevifyPython
fidecastro/jevifyPython164GitHub stars
End-to-end toolkit that synthesizes decision data, trains and calibrates candidate scorers, evaluates fixed splits, and serves Jev-compatible inference for interactive web decisions.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
zwliJay/jev-forgePython
zwliJay/jev-forgePythonFrom 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.