User repository on GitHub
a free multi-modal jev API for everyone
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Demos & Experiments
Amigos Jev
Free Jev-style multimodal decision API that takes images plus state and returns a choice, score, or yes-no answer with probabilities in a few hundred milliseconds.
- Stars
- 6
- Published by
- Community
- Author
- amigos-robot
- Last verified
- Oct 2026
Works forJev for Classification
Used inJev for Software & DevTools
Highlights
- Accepts text plus up to 4 images with native multimodal scoring, not caption-then-judge.
- Answers in 100-500ms with one forward pass per question for control-loop timing.
- Returns per-option probabilities with confidence for thresholding or human handoff.
- Offers a hosted systemone endpoint plus live snake and sketch demos without signup.
Watch out
A third-party Jev-style service with its own weights and endpoint, not the TypeSafe hosted model, so behavior and availability follow its operator. 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.
No. It reuses the next-token distribution as the decision output without generating text, with probabilities renormalized inside your option set. Outputs stay inside the actions you define, which is why it fits hardware control loops.
Send up to 4 images with text or JSON state to the systemone endpoint. Each question costs one forward pass, so camera-in plus action-out loops and per-frame demos like the snake stream are the intended shape.
The listing says no signup and free calls through the hosted endpoint, with docs and interactive demos on the site. Treat quotas, uptime, and model changes as operator-controlled since this is not the official API.
Use the probability and confidence per option to stop or escalate when uncertain. The action space is yours to define, so low-confidence handling stays in your code rather than in the model.
POST to https://jev.amigos-robot.com/v1/systemone with a state and typed questions; it is free and needs no application. It accepts text, JSON, and up to 4 images, and returns a probability and confidence per option.
It is a Jev-style native multimodal decision model with about 20B active weights. It uses the multimodal model's next-token probability distribution directly as the decision output rather than generating text.
The action space is defined by you and the model can only choose among the given options, so it cannot emit an unplanned instruction. Each action carries a probability, and low confidence lets you stop the robot or hand it to a human.
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
Community Spring Boot starter that injects a JevClient into Spring MVC apps to ask typed noul, choice, and score questions against the TypeSafe API.
Jev Demos & Experiments#community#awesomejev#catalog-import
Community
danvega/jev-spring-boot-starterJava
danvega/jev-spring-boot-starterJava39GitHub 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/jevifyPythonFrom 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.