trycatchkamal/typesafe-jev-traffic-demo
User repository on GitHub
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Demos & Experiments
Jev at the Junction
A local proof of concept where Jev picks which of two conflicting signal phases gets the green light at a real Hong Kong intersection, using near real-time keyless open-data traffic sensors. A phase-timing state machine enforces minimum and maximum green, amber and all-red around the model's answer.
- Stars
- 1
- Language
- Python
- Reading time
- 3 min
- Published by
- Community
- Author
- trycatchkamal
- Last verified
- Oct 2026
- Languages
- Python
- Capabilities
- Open sourceRuns locallyRequires API key
Highlights
- Polls a free, keyless near real-time open-data feed at four sensors: Kwun Tong Bypass crossing Kai Fuk Road in Hong Kong.
- Every ~30 seconds it asks Jev five questions: congestion per side, which side needs priority, whether the gap warrants action, and data reliability.
- A phase-timing state machine runs around Jev's answer — the model proposes, the state machine enforces minimum green, amber and all-red clearance.
- Minimum green is adaptive: it shrinks toward a 10-second floor when one side is clearly worse off, and stays at the default otherwise.
- The README is explicit that it is a simulation: Hong Kong publishes a read-only sensor feed and no write API for traffic signals.
Quickstart
bash
pip install -r requirements.txt
cp .env.example .env
python junction.pyWatch out
Copy reviewed against the repo README on 2026-10-15; the README's endpoint table is truncated in the digest and no accuracy or timing benchmarks for the Jev calls are published. It states clearly that the project cannot control any real traffic signal, since Hong Kong's Transport Department exposes only a read-only feed.
Questions about this entry
Short answers drawn from the linked project or page.
A local proof of concept that has Jev — TypeSafe's typed decision model — pick which of two conflicting signal phases gets the green light at a real Hong Kong intersection, using near real-time open-data traffic sensors from a free, keyless feed.
Every ~30 seconds the script sends the current sensor state and asks five questions: how congested each side is, which one needs priority, whether the gap between them is large enough to act on, and how reliable the data looks. A state machine then enforces the safety-critical timing.
No. This is a simulation. Hong Kong's Transport Department publishes no write API for signals, only a read-only sensor feed, and everything downstream — phase timing, amber, all-red clearance, safety limits — runs in this process's own memory.
Install with pip install -r requirements.txt, copy .env.example to .env and set JEV_API_KEY, then run python junction.py and open http://localhost:8790 for the dashboard with the live decision and manual controls.
Defaults include a 40-second minimum green (adaptive down to a 10-second floor), a hard 120-second maximum that forces a switch, 4 seconds of amber, 2 seconds of all-red clearance, and a 0.6 minimum switch confidence before acting on Jev's recommendation.
Yes. Set JEV_URL and JEV_MODEL in .env if your Jev access is via a gateway; the defaults are https://api.typesafe.ai/v1/systemone and jev-latest. The script loads .env automatically and real shell environment variables take priority over it.
It shows the live decision and a plain-English readout of Jev's reasoning, and it offers manual controls to inject extra simulated traffic on either side or trigger an emergency-vehicle preemption hold.
More Jev resources like this
A no-dependency single-page classifier that unpacks a Jev call: paste an API key, edit a choice question, and see the request JSON, per-category probabilities, confidence, latency, token usage and the equivalent curl.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
q93304989-bit/jev-labHTML
q93304989-bit/jev-labHTMLA workshop of small, inspectable Python experiments for TypeSafe System One models: ten critical bays, from fraud screening to emergency dispatch, each giving Jev a realistic state while deterministic policy keeps authority over the next action.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
jgalego/jevs-garagePython
jgalego/jevs-garagePythonBrowser-only Laya demo suite: four Vite apps — a Traditional Chinese decision playground, Gomoku, Big Two and a 3D maze — run the Laya ONNX model on-device through Transformers.js, downloading fp16 weights from Hugging Face on first use.
Jev Demos & Experiments#community#catalog-import#madewithjev
Community
killkli/open-jev-layaPython
killkli/open-jev-layaPythonTraffic signal control with Jev applied to SUMO simulation: each control step packs the road network into one state with a choice question per intersection, and the model's chosen phase, distribution and confidence drive the signal with local-ranking fallbacks.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
usail-hkust/jevlightPython
usail-hkust/jevlightPythonAn autonomous ViZDoom agent that decouples Doom's simulation from Jev SystemOne inference: the game runs at 35 ticks/sec while Jev is queried asynchronously at about 10 Hz for macro goals, targets, movement and firing decisions.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
amoghcreator/doom-jevPython
amoghcreator/doom-jevPythonA 12,000-neuron fly-brain subgraph of the maleCNS connectome fights Jev in mk.js: the fly simulates spikes in the browser at 10 decisions per second while Jev reads the fight as JSON and answers one typed choice among seven moves.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
lavallee/mk-jev-fly-brainJavaScript
lavallee/mk-jev-fly-brainJavaScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.