Job Risk Analyzer

A TypeScript/Node.js CLI and REST API that uses TypeSafe AI's Jev System One model to evaluate an occupation's exposure to AI-driven layoffs, likely future trajectory, need for human accountability, and overall resilience.

The project works immediately without credentials. When JEV_API_KEY is absent, it logs a warning and uses deterministic, occupation-aware mock data with the same validated shape as a live Jev response.

Requirements

• Node.js 20 or newer
• npm
• Optional: a TypeSafe AI API key for live analysis

Setup

npm install
cp .env.example .env

Environment variables are read from the process. Export the key before running the app:

export JEV_API_KEY="your-api-key"

If the variable is unset or blank, the CLI and server automatically use mock evaluations.

CLI

Run through the npm script:

npm run cli -- --title "Data Entry Clerk"
npm run cli -- --title "Cybersecurity Analyst" --desc "Protects cloud infrastructure and responds to incidents"

Or invoke tsx directly:

npx tsx src/cli/index.ts --title "Data Entry Clerk"

The table reports layoff risk, future trajectory, trajectory confidence, human-accountability probability, and a derived resilience score.

REST API

Start the server:

npm run server

Development mode restarts when TypeScript files change:

npm run dev

The server listens on 0.0.0.0:3000 by default. Override it with HOST and PORT.

Open http://localhost:3000 in a browser to use the responsive web interface. Enter an occupation and optional role context, then select Run occupation forecast. The page calls the REST API from the browser and displays the risk forecast in the report panel.

Analyze an occupation:

curl -X POST http://localhost:3000/api/analyze \
  -H "Content-Type: application/json" \
  -d '{"title":"Registered Nurse","description":"Provides bedside care in an acute hospital"}'

Health check:

curl http://localhost:3000/health

Deploy to Vercel

Vercel detects the Fastify entrypoint in src/index.ts without additional configuration:

npx vercel deploy --prod

Add JEV_API_KEY in the Vercel project's environment variables to use live Jev analysis. Without it, the deployed API uses the typed mock fallback.

Scripts

Script	Purpose
npm run dev	Run the API server in watch mode
npm run cli -- --title "..."	Run an occupation analysis
npm run server	Start the API server
npm run typecheck	Run strict TypeScript checks

How It Works

The analyzer sends three independent questions in one POST https://api.typesafe.ai/v1/systemone request using model jev-latest:

• layoff_risk: a five-level Score normalized to 0..1 in application code
• future_trajectory: a Choice among high_layoff_risk, transformed_hybrid, and ruling_trend
• human_accountability: a Noul probability that a human must remain accountable for consequential outcomes

The resilience score is a transparent application-level calculation combining inverse layoff risk, human accountability, and a small trajectory adjustment. Results are forecasts, not employment or financial advice.

Example Response

{
  "title": "Data Entry Clerk",
  "layoffRisk": 0.82,
  "futureTrajectory": "high_layoff_risk",
  "trajectoryConfidence": 0.72,
  "humanAccountability": 0.28,
  "resilienceScore": 0.115,
  "source": "mock",
  "model": "jev-latest-mock"
}