vignu10/parking-decision-layer
User repository on GitHub
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Demos & Experiments
Parking Operations Decision Layer
A pilot build of the Parking Operations Decision Layer PRD: the support desk, agent guard and shared decision layer send each item to Jev as typed questions in one call, then rules and a confidence gate route it to AUTO, REVIEW, HUMAN or BLOCKED with an audit log.
- Stars
- 0
- Language
- TypeScript
- Reading time
- 2 min
- Published by
- Community
- Author
- vignu10
- Last verified
- Oct 2026
- Languages
- TypeScript
- Platforms
- Vercel AI Gateway
- Capabilities
- Open sourceRuns locally
Highlights
- Flow: item to Jev typed questions in one call, then business rules, then AUTO, REVIEW, HUMAN or BLOCKED, then actions and an audit log.
- Rules are layered: hard rules for safety, money and policy first, then a confidence gate equal to the weakest gating question.
- Runs fully offline with a deterministic mock Jev; JEV_PROVIDER=mock is the default, so demos never depend on the network.
- Real providers: vercel with AI_GATEWAY_API_KEY (Vercel AI Gateway, typesafe-ai/jev, no waitlist) or typesafe with TYPESAFE_API_KEY.
- Scope spans support desk S-1 to S-8, agent guard D-9 and D-10, and shared layer F-1 to F-8; npm test runs 38 tests.
Quickstart
bash
npm install
cp .env.example .env.local
npm run dev
npm testWatch out
README reviewed 2026-10-17; the captured text cuts off inside the typesafe provider row, and this is a pilot build of an internal PRD with 0 stars and no license. The offline mode is deterministic keyword heuristics, not measured model behavior.
Questions about this entry
Short answers drawn from the linked project or page.
A pilot build of the Parking Operations Decision Layer PRD covering the support desk (S-1 to S-8), the agent guard (D-9, D-10) and the shared decision layer (F-1 to F-8).
Each item goes to Jev once with typed questions; the result then passes hard rules and a confidence gate before routing to AUTO, REVIEW, HUMAN or BLOCKED, with actions and an audit log.
Yes. JEV_PROVIDER=mock is the default and runs fully offline with deterministic keyword heuristics, so demos never depend on the network.
Set JEV_PROVIDER=vercel with AI_GATEWAY_API_KEY for the Vercel AI Gateway (typesafe-ai/jev, no waitlist), or JEV_PROVIDER=typesafe with TYPESAFE_API_KEY against api.typesafe.ai.
It uses the weakest gating question's confidence, applied after the hard rules that cover safety, money and policy, according to the README's flow diagram.
npm test runs 38 tests, including the PRD section 12 acceptance criteria, per the README's run instructions.
It is a pilot build of an internal PRD with 0 stars and no license; the captured README cuts off inside the provider table, and its stated principle is that AI sorts and suggests while rules and people keep control.
More Jev resources like this
Three TypeScript demos built on one pattern for students: Jev answers a typed question with a probability, your code decides what to do, and Claude works only where it is needed. Demos cover Jira triage, pre-flight checks and PR routing.
Jev Demos & Experiments#community#github#demo
Community
soyjorgepilo/jev-demoTypeScript
soyjorgepilo/jev-demoTypeScriptA TypeScript starter for Jev called through the Vercel AI Gateway with the AI SDK. It wraps evaluate calls in an askJev helper and ships a support-ticket triage example using choice, score and boolean questions.
Jev Demos & Experiments#community#github#typescript
Community
1leetadmin/JEVTypeScript
1leetadmin/JEVTypeScriptA Chinese-language e-commerce customer-service demo on FastAPI, LangGraph and React: adaptive retrieval routing with hybrid BM25-plus-vector search, and intent classification switchable between TypeSafe Jev's structured choice interface and DeepSeek.
Jev Demos & Experiments#community#github#python
Community
gbh1234/Adaptive-ragPython
gbh1234/Adaptive-ragPythonA per-turn model router for the Pi coding agent: a small judge picks the cheap or the strong tier for each message, with multi-model failover and task-level orchestration, plus an optional Jev judge that answers with a calibrated probability.
Jev Tools & CLIs#community#github#router
A demo that has three decision models read recent news and price Polymarket and Kalshi questions, then compares their probabilities with live market odds and replays how each model changed its mind article by article.
Jev Demos & Experiments#community#github#demo
Community
yorkeccak/model-vs-marketTypeScript
yorkeccak/model-vs-marketTypeScriptVercel Labs example that routes three forms by context with Jev and AI SDK: Jev's team choice is accepted when its confidence is at least 95%, otherwise openai/gpt-6-luna-fast independently evaluates the submission and its choice becomes final.
Jev Demos & Experiments#community#catalog-import#logicrw
From 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.