I built a FREE tool that uses AI (Hidden Markov Models) to analyze the "hidden regime" of any stock

Discussion in 'App Development' started by Arse, Aug 25, 2025 at 6:31 AM.

  1. Arse

    Arse

    Hey everyone!

    I'm a dev with a passion for quantitative finance, and I've spent the last few months building a web tool I wanted to share with you all to get some feedback.

    The core idea: Price charts tell you what happened, but not always why. I wanted to try and understand the underlying "regime" or "mood" of the market for a specific asset and, most importantly, measure how reliable that prediction is.
    Imagine the market has a hidden mood (Positive Momentum, Negative Momentum, or Consolidation). We can't see it directly, but we can see its effects: the daily price movements. My app uses an AI model (a Hidden Markov Model or HMM) that analyzes these movements to figure out the market's most likely "mood" right now.
    What can you do with it?

    • Single Analysis: Enter any ticker (AAPL, GOOG, BTC-USD...) and it tells you the dominant trend, transition probabilities, and crucially, how reliable the model has been for that specific asset in the past (using a rigorous Walk-Forward backtest).

    • AI Analysis (Gemini): So you don't drown in data, an AI generates a qualitative summary that explains what all the numbers mean, just like a financial analyst would.

    • ⚖️ AI Comparator: Can't decide between two stocks? Pit them head-to-head, and the AI will give you a verdict on which one looks more favorable based on the model's data.

    • Portfolio Manager + AI Advisor: Build your own portfolio, and the AI will generate a strategic report identifying your strongest assets, those that need reviewing, and giving you rebalancing suggestions.

    • Index Screener: Scan entire indices (S&P 100, Nasdaq 100, DAX, etc.) at once to find potential buy/sell candidates based on strict risk/return criteria.
    It's 100% free, runs directly in your browser, and requires no sign-up.

    Link: https://stockstrend.app

    Tech Stack (for the nerds): React, TypeScript, TailwindCSS, Recharts. The entire HMM analysis runs client-side in pure TS. The AI part is powered by the Google Gemini API.

    I'd love for you to try it out and give me your feedback. Is it useful? What's missing? Did you find any bugs? Any and all feedback is gold!

    Disclaimer: This is a purely educational tool, not financial advice. Always do your own research.
     
    MarkBrown likes this.
  2. MarkBrown

    MarkBrown

    get rid of this please it has no use at all - it cheapens what your trying to do

    MACD Indicator

    also you don't have the SP500?
     
    Last edited: Aug 25, 2025 at 10:13 AM
  3. If you are going to have static options for indices, scan them in the background in advance regularly, let's say every 5 minutes. So you can display the result immediately. I got bored of waiting for NQ to finish the scan.

    Think about that you are competing with other scanners that are also free, so the website must be quicker.
     
    MarkBrown likes this.
  4. MarkBrown

    MarkBrown


    agreed it should be alive - also i need more clear - this is a buy this is a sell

    if you can solidify what your model's commitment is then you have just another pretty site.

    ps i'm going to borrow that icon for the light and dark color scheme and its code.
     
  5. Also, test it thoroughly before sending it to the public.

    upload_2025-8-25_16-27-2.png
     
  6. Arse

    Arse

    Thank you very much for your time testing the application, and especially for your comments and suggested improvements
     
    Drawdown Addict likes this.
  7. Arse

    Arse

    The purpose of labeling the results of the Index Analysis as Possible Buys or Possible Sells is to signal the need for a deeper analysis of the ticker, either individually or as part of a portfolio. Thanks for your comments
     
  8. MarkBrown

    MarkBrown

    so its vauge like everyone else has

    be like me its either a buy or a sell no vagueness at all

    OddBall.Systems | E-mini S&P 500 Futures Day Trading
     
  9. nz_melon

    nz_melon

    Was that built with svelte? What chart/ui component library did you use?