Rager's Get Rich Slow Scheme

Automated Kalshi sports prediction market scanner
_{Buy YES contracts at 88–99¢ on games that are already decided. Collect $1 at settlement.}

A project by Brent Rager — founder of Smoo AI AI agents, dev tools, CRM, support & campaigns that integrate with everything you build. Let's talk.

---

How It Works

The scanner watches live sports games across NBA, NHL, MLB, NFL, MLS, Premier League, La Liga, NCAA, and UFC — and buys YES contracts on Kalshi when:

1. ESPN confirms the game is in its final minutes — 4th quarter <5min, 9th inning, final period, etc.
2. The leading team has a comfortable margin — 8+ pts NBA, 2+ goals NHL/soccer, 10+ pts NFL, etc.
3. Kalshi YES price is 88–99¢ — market already expects this outcome
4. Sufficient liquidity — 50+ volume, active bid

The edge: Kalshi prices lag behind live game state. A team up 15 points with 2 minutes left in the 4th quarter is a near-certainty, but the YES contract might still be at 92¢. We buy at 92¢, collect $1 at settlement. Small margins, high win rate.

Architecture

graph TD
    ESPN["🏈 ESPN API<br/>scores, periods, clocks · 10s poll"]
    KWS["⚡ Kalshi WebSocket<br/>real-time prices + settlements"]
    KREST["📡 Kalshi REST API<br/>discover markets, place orders · 5s"]
    SCANNER["🐍 Scanner<br/>Python / asyncio"]
    DB["🗄️ SQLite on EFS<br/>trades, balance, opportunities"]
    API["🚀 FastAPI<br/>/api/*"]
    DASH["🖥️ Next.js Dashboard<br/>getrich.rager.tech"]

    ESPN --> SCANNER
    KWS --> SCANNER
    SCANNER --> DB
    SCANNER --> KREST
    DB --> API
    API --> DASH

    style ESPN fill:#1a1a2e,stroke:#FF0000,color:#fff
    style KWS fill:#1a1a2e,stroke:#64748b,color:#fff
    style KREST fill:#1a1a2e,stroke:#64748b,color:#fff
    style SCANNER fill:#1a1a2e,stroke:#F59E0B,color:#fff
    style DB fill:#1a1a2e,stroke:#003B57,color:#fff
    style API fill:#1a1a2e,stroke:#009688,color:#fff
    style DASH fill:#1a1a2e,stroke:#fff,color:#fff

Loading

Real-Time Data Pipeline

Layer	Source	Method	Purpose
Prices	Kalshi WebSocket	ticker channel, real-time	Live YES/NO bid-ask streaming
Settlements	Kalshi WebSocket	market_lifecycle_v2	Instant win/loss detection
Markets	Kalshi REST API	Poll every 5s	Discover new markets, place orders
Game State	ESPN API	Poll every 10s	Score, period, clock verification
Dashboard	FastAPI → Next.js	REST	P&L tracking, live games, analytics

What-If Strategy Tracking

Five shadow strategies run in parallel — evaluating every market against different price thresholds, timing windows, and lead requirements. Each tracks hypothetical P&L so we can backtest parameter changes before risking real capital.

Tech Stack

Python 3.12	FastAPI	SQLite	Next.js 16	React 19	Tailwind CSS
Docker	AWS ECS	SST v3	Cloudflare	WebSockets	ESPN API

Dashboard

The dashboard at getrich.rager.tech shows:

• Account Value — step chart tracking balance over time with each win/loss
• Live Games — real-time ESPN game state with Kalshi market prices, color-coded by betting criteria
• Scanner Config — live view of all trading parameters and per-sport rules
• Trade History — every bet placed with P&L
• Strategy Comparison — side-by-side what-if analysis across 5 parameter sets
• Stats — win rate, realized P&L, open positions

Deployment

Prerequisites

• SST v3 + pnpm
• uv (Python package manager)
• AWS account (us-east-2)
• Cloudflare account (DNS)
• Kalshi API keys

Secrets

SST secrets must be set before first deploy:

npx sst secret set KalshiApiKey "your-kalshi-api-key-id"
npx sst secret set KalshiPrivateKey "-----BEGIN RSA PRIVATE KEY-----\n..."
npx sst secret set DashboardPassword "your-dashboard-password"
npx sst secret set ApiToken "your-api-bearer-token"

Infrastructure

SST provisions everything in sst.config.ts:

Resource	What	Notes
VPC	Networking with NAT	EC2-based NAT to save cost
ECS Cluster	Container orchestration	Single service runs API + Scanner
EFS	Persistent storage	SQLite database lives here
S3 Bucket	DB backups	Periodic SQLite snapshots
Next.js (SST)	Dashboard	Lambda + CloudFront via OpenNext
Cloudflare DNS	Domain routing	getrich.rager.tech + getrich-api.rager.tech

Commands

# Deploy to AWS (requires AWS creds + Cloudflare token in .envrc)
pnpm sst:deploy

# Local development (Docker Compose)
pnpm dev

# Run API locally (without Docker)
pnpm dev:api

# Tear down all infrastructure
pnpm sst:remove

CLI

A react-ink TUI for managing the scanner from the terminal:

# View current config (interactive TUI)
pnpm cli config

# Update a config value
pnpm cli config set trading.min_yes_price 90

# View stats / trades
pnpm cli stats
pnpm cli trades

# JSON output for piping / automation
pnpm cli stats --json
pnpm cli trades --json | jq '.trades[0]'

Requires API_TOKEN env var (or --token flag). Set GETRICH_API_URL to override the default API endpoint.

Runtime Configuration (API)

The scanner reads config from the database at startup. Update settings via the CLI above or the API directly:

# Update a config value (requires API_TOKEN)
curl -X PUT https://getrich-api.rager.tech/api/config \
  -H "Authorization: Bearer $API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"key": "trading.min_yes_price", "value": "92"}'

# View current config
curl https://getrich-api.rager.tech/api/config \
  -H "Authorization: Bearer $API_TOKEN"

Development

# Format + lint Python
uv run ruff format . && uv run ruff check . && uv run ty check

# Format + lint TypeScript
cd dashboard && pnpm oxfmt . && pnpm oxlint

# Build dashboard
cd dashboard && pnpm build

A pre-commit hook runs automatically: formats Python (ruff) and TypeScript (oxfmt), then type-checks with ty.

---

_{Built by Brent Rager at Smoo AI with Claude Code}