typesafe-playground
Interactive Jev experiments, from a support-message decision lab to 3D self-driving demos with steer, speed and overtake decisions.
Overview
This repository collects small interactive experiments with Jev that make inputs, judgments and resulting behavior visible, including hesitation and mistakes. Demo01, Decision Lab, turns a support message into six independent judgments (owning team, human request, refund request, urgency, frustration and reproduction steps) that code combines into a routing preview.
The driving demos have Jev choose lane and target speed for a car on a two-way road with random traffic, overtaking, oncoming cars, pedestrian crossings and remembered speed signs, with camera views, sensor overlays and a decision inspector. Demo03 is the original implementation used in a published post, and Demo04 restyles it. The author describes these as experiments with real API calls, not benchmarks.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which team owns this support message? | — |
| Yes / no | Is the customer requesting a refund / a human? | yes · no |
| Choice | Which lane and target speed should the car take? | — |
How it works
- Choose a demo and start it locally.
- Structured state (message or sensors) is sent to Jev.
- Code combines answers into routing or driving actions.
- Inspectors show inputs, judgments and behavior.
Strengths
- Makes model decisions and mistakes visible.
- Covers both text routing and continuous control.
Caveats
- Experiments, not benchmarks.
- Routing is a preview; no real tickets or refunds.
At a glance
- Type
- App
- Status
- prototype
- Runtime
- local web apps
- Has code
- yes
Use cases and integrations
How curated lists describe it
Interactive playground: small Jev experiments that put the decision on screen, from routing a support message to steering a car in a 3D world.
yibie/awesome-jevInteractive Jev experiments for support-routing previews and 3D driving simulations.
logicrw/awesome-jev-projectsInteractive experiments from support routing to a 3D driving simulation: structured sensor state in, typed steer, brake, and overtake decisions out.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide in the driving demo?
Lane and target speed, based on camera, radar and sign state.
Does Decision Lab send real tickets?
No, routing is only a preview.
Are these benchmarks?
No, they are interactive experiments.
Related projects
Korean RL tycoon game training a company decision model that answers situations with choice, probabilities and confidence.
ProjectprototypePythonToy archipelago air-traffic sim where each plane asks Jev every 1.6 s whether to divert, declare an emergency, give way or land.
DemoprototypeTypeScriptSmall Python demos of Jev: parallel Noul ticket checks with measured speedup, and a three-way Choice citation and claim checker.
AppprototypePythonTwo beginner Python apps using Jev: a message sorter and a support ticket router that send uncertain items to human review.
ProjectprototypePythonPython sandbox turning Jev issue-category and reproduction judgments into testable routing proposals, with offline replay.
ProjectprototypePythonWorked Jev examples for ticket triage and reranking, runnable through OpenRouter without an early-access key, with offline tests.
ProjectprototypePython