The Pit

Gamified scenario training & MTSS agent for options trading.

Production

• URL: https://thepit.up.railway.app/

Quick Start (Local)

docker-compose up --build

• Frontend: http://localhost:5173
• Backend: http://localhost:8000
• API Docs: http://localhost:8000/docs

Deploy to Railway

1. Install the Railway CLI and log in:

 npm install -g @railway/cli
 railway login

2. Create a new project and add services:

 railway init

3. Add PostgreSQL and Redis add-ons from the Railway dashboard.
4. Create two services pointing to this repo:

• Backend — root directory: backend/
• Frontend — root directory: frontend/

5. Set environment variables on the Backend service:

 ANTHROPIC_API_KEY=sk-...
 OPENAI_API_KEY=sk-...
 JWT_SECRET_KEY=<generate-a-strong-secret>
 FRONTEND_URL=https://<frontend-service>.up.railway.app
 CORS_ORIGINS=https://<frontend-service>.up.railway.app
 APP_ENV=production

DATABASE_URL, REDIS_URL, and PORT are auto-injected by Railway add-ons.

6. Set environment variables on the Frontend service:

 VITE_API_URL=https://<backend-service>.up.railway.app

7. Deploy:

 railway up

8. Seed the database (first deploy only):

 railway run -s backend bash -c "SEED_PROD=true python -m app.seed"

Staging Environment

Staging mirrors production but uses a separate Railway project and database so you can test changes before going live.

Local staging (Docker Compose)

docker compose -f docker-compose.yml -f docker-compose.staging.yml up --build

This uses a separate Postgres volume (pgdata_staging) so staging data stays isolated from local dev.

Deploy staging to Railway

Follow the same "Deploy to Railway" steps above but with these differences:

1. Create a separate Railway project (e.g. "the-pit-staging").
2. Set APP_ENV=staging on the Backend service.
3. Pass BUILD_ENV=staging and VITE_API_URL=https://<staging-backend>.up.railway.app/api as build args on the Frontend service.
4. Point the Railway project at the staging-environment branch (or whichever branch you use for staging).

The staging backend runs without --reload (like production) but keeps seed data enabled by default so you always have test accounts available.

Test Accounts

Seed the database with python -m app.seed from the backend directory. By default, development test users are created; set SEED_PROD=true for demo/admin accounts. Add SEED_CLEAN=true to move non-seed users out of demo orgs for a clean demo slate.

For multi-tenant logins, account resolution follows org context. On hosted UI:

• thepit.up.railway.app resolves to org slug thepit
• acme.yourapp.com resolves to org slug acme (or pass org_slug to /api/auth/login)
• If you use a single shared URL (no custom subdomains), choose the org in the login page selector (thepit or acme).

Development (default)

Org	Email	Password	Role	Display Name
thepit	trader@thepit.dev	trader123	analyst	Test Trader
thepit	alex@thepit.dev	alex123	analyst	Alex Chen
thepit	maria@thepit.dev	maria123	analyst	Maria Santos
thepit	james@thepit.dev	james123	intern	James Kim
thepit	priya@thepit.dev	priya123	analyst	Priya Patel (most advanced)
thepit	admin@peak6.com	peak62026	org_admin	Peak6 Super Admin
acme	admin@acme.dev	acme2026	org_admin	Acme Admin
acme	analyst1@acme.dev	acme2026	analyst	Acme Analyst One
acme	associate@acme.dev	acme2026	associate	Acme Associate
acme	trainer@acme.dev	acme2026	trainer	Acme Trainer
acme	intern@acme.dev	acme2026	intern	Acme Intern

Production (SEED_PROD=true)

Org	Email	Password	Role	Display Name
thepit	demo@thepit.dev	demo2026	analyst	Demo Trader
thepit	advanced@thepit.dev	advanced2026	analyst	Advanced Demo
thepit	trader2@thepit.dev	demo2026	analyst	The Pit Trader Two
thepit	associate@thepit.dev	demo2026	associate	The Pit Associate
thepit	intern@thepit.dev	demo2026	intern	The Pit Intern
thepit	admin@peak6.com	peak62026	org_admin	Peak6 Super Admin
acme	admin@acme.dev	acme2026	org_admin	Acme Admin
acme	analyst1@acme.dev	acme2026	analyst	Acme Analyst One
acme	associate@acme.dev	acme2026	associate	Acme Associate
acme	trainer@acme.dev	acme2026	trainer	Acme Trainer

Testing

Backend (pytest)

cd backend && uv run pytest tests/ -v

Tests: auth, constants, grading agent, leaderboard, MCQ generation/grading/pool, progression, RAG, scenario engine.

Golden set checks

cd backend && uv run python evals/run_golden_checks.py

Golden sets live under backend/evals/golden/ and cover scenario prompt construction plus JSON output parsing invariants.

Frontend (Vitest)

cd frontend && npm test -- --run

(Use --reporter=verbose for detailed output.)

Tests: AuthContext, MCQCard, MCQFeedback, QuickFirePage, TrainingPage (prefetch).