▲ Open source · runs entirely in your browser

Research Any Stock, Forecast It with Kronos AI

Quality scores, a 2-click DCF, analyst targets and institutional ownership — plus 30-day price forecasts sampled from an open-source foundation model for candlesticks.

0
Companies Covered
0.0M
Kronos AI Parameters
0
Sampled Forecast Paths
0
Backend Servers
Foundation Model

30-Day Forecasts from Kronos AI

Kronos AI is an open-source decoder-only foundation model pre-trained on K-lines from 45+ exchanges. Gnzub runs it over real daily bars and samples eight possible futures per company, so you see a range of outcomes instead of a single confident line.

  • Median path with a 10–90% uncertainty band
  • Every sampled path shown, not just the average
  • Model card with the exact context, horizon and sampling settings

Open the forecast  →

Valuation Tool

2-Click DCF Calculator

Run a full Discounted Cash Flow valuation on any stock in seconds. Our calculator auto-populates financial data, calculates WACC, and projects 5-year free cash flows — giving you an intrinsic value estimate instantly.

  • Auto-populated from real financial statements
  • Adjustable assumptions with live recalculation
  • 5-year projection table with terminal value

Open the calculator  →

Wall Street

Aggregated Analyst Price Targets

See where Wall Street thinks a stock is heading. We aggregate price targets from dozens of analysts to show you the consensus, the outliers, and the distribution — all in one view.

  • Average, highest, and lowest analyst targets
  • Target distribution histogram
  • Upside/downside potential vs. current price

See price targets  →

Smart Money

Follow the Smart Money

Track what the biggest institutional investors and company insiders are buying and selling, from their own 13F and Form 4 filings. See who holds the most and how positions moved.

  • Top institutional holders with ownership %
  • Institutions adding, trimming, opening and closing positions
  • Insider transactions as filed on Form 4

Explore ownership  →

How Gnzub Works

A static site, a foundation model, and no server in between.

1

Kronos AI runs offline

A Python script pulls real daily OHLCV, runs the model on your own machine and writes the sampled paths to a JSON file.

2

The site reads that file

Charts, scores, the DCF and the screener are plain HTML and vanilla JavaScript — no build step, no backend, no tracking.

3

You add a key, or not

Drop in a free API key for live prices, or leave it off and work from the committed snapshot.

What's Actually in Here

Straight about which numbers are real and which are placeholders.

Real

Prices, daily closes, reported revenue and net income, cash, debt, analyst price targets and rating spreads — all imported from market data. The Kronos AI forecasts, the DCF and the quality scores are computed from them.

Filings

Institutional holders come from 13F filings and insider transactions from Form 4, via Nasdaq. News headlines are sample text wherever no news feed is connected, and are labelled as such.

Not advice

Kronos AI samples plausible futures; it does not know what a stock will do. Nothing here is a recommendation, and the forecasts are a snapshot from one model run, not a live signal.

Run It Yourself

Clone the repo, generate a fresh Kronos AI snapshot on your own GPU, push it to GitHub Pages. No account, no backend, no tracking.

Open the dashboard
Static site · vanilla JavaScript · your API keys stay in your browser