Settings

Manage your preferences

Account

Invite-only — enter your code to use Alpha

Checking account…

Appearance

Light mode

System Status

Backend + data engine health, with one-tap wake

BackendNot checked
Data engine (ML)Not checked

Free-tier services sleep after ~15 min idle — this wakes them on demand. A cold boot takes ~30–60s.

How Alpha's AI works

The ML pipeline behind every score

For every stock, four agents rate it on a 0–10 scale. The combined score drives the call — ≥7 Strong Buy · 5–7 Hold · <5 Avoid.

Fundamental Agent

P/E valuation, ROE, debt/equity, revenue growth, institutional holding.

Technical Agent

RSI-14, MACD crossover, and 20-day annualised volatility.

Sentiment Agent

VADER reads news headlines — Yahoo Finance, Google News RSS fallback — and labels them bullish / bearish / neutral.

Risk Critic

Annualised volatility above 30% cuts the score by 10%; sets the risk score and position-size multiplier.

Runs on a separate FastAPI microservice fetching Yahoo Finance data, disk-cached to dodge rate limits. Screens cover a 20-stock ranked universe and 100+ discovery tickers. Orchestration runs directly — LangGraph isn't wired up yet.

About

Versionv0.1.0
Stack
Next.js + NestJS + FastAPI

Alpha uses AI-powered multi-agent scoring to analyse NSE stocks. All data sourced from Yahoo Finance. Not financial advice.