MarginCall

Alerts on company fundamentals and market data. Pick a stock, choose a metric, set a rule, and MarginCall checks it against live Yahoo Finance data every minute and records every time it fires.

What it does

• Three kinds of signal.
  • Metric vs metric: "Apple's total debt is above its total cash".
  • Metric vs value: "current ratio below 1".
  • Metric vs % change: "Tesla falls 15% from the price when I set this".
• Five operators: >, <, =, >=, <=.
• 27 metrics across the balance sheet, cash flow, financial ratios, the income statement and market data, each shown with its live value while you build the signal.
• Every-minute monitoring. Each active signal is re-evaluated against fresh data, and every trigger is stored with the exact values that caused it.
• Accounts. Sign-up and login with JWT, and each user sees and manages only their own signals, with a switch to pause or resume any one.

Build a signal from live data	Trigger history

How it works

flowchart LR
  UI[Next.js dashboard] -- JWT --> API[NestJS API]
  API <--> DB[(PostgreSQL)]
  API -- metrics for a ticker,<br/>initial values --> PY[Python service]
  PY -- every minute: latest data<br/>for active signals --> API
  PY <--> YF[Yahoo Finance]

Loading

1. Build. The dashboard asks the API which metrics exist for a ticker, and the Python service looks them up on Yahoo Finance with their current values. The user picks a metric, an operator, and a second metric, a fixed value or a percentage.
2. Save. The API stores the signal with the starting values of its metrics, so percentage rules have a baseline.
3. Watch. Every minute, the Python service finds the signals due for an update, fetches their metrics, and posts the data to the API.
4. Evaluate. The API compares the values according to the signal type and operator. When a rule holds, it marks the signal triggered and writes a history record with the values and a timestamp.

API

Method	Path
POST	/auth/register, /auth/login	Create an account or sign in; returns a JWT
POST	/signals	Create a signal
GET	/signals/my	The signed-in user's signals
PUT	/signals/:id/toggle	Pause or resume a signal
GET	/signals/:id/history	Every trigger, with the values at the time
GET	/signals/available-metrics/:ticker	Metrics for a ticker, grouped, with live values
GET	/signals/operators, /signals/types	Supported operators and signal types
POST	/signals/:id/update, /signals/:id/trigger	Used by the Python service to deliver fresh data

The Python service exposes /health, /available-fields/:ticker and /fetch-initial-values.

Run it

docker compose up -d --build              # PostgreSQL, NestJS API (:3000), Python service (:5001)

cd frontend && npm install
NEXT_PUBLIC_API_URL=http://localhost:3000 npx next dev -p 3001   # dashboard on :3001

Tech

• Frontend: Next.js 14, TypeScript, Tailwind CSS, Axios
• API: NestJS, TypeORM, PostgreSQL 15, Passport JWT, bcrypt
• Data service: Python, Flask, yahooquery, schedule
• Infrastructure: Docker Compose