Inspiration
Most stock research tools either overwhelm users with raw financial data or ask them to trust a black box AI answer. We wanted a clearer way to evaluate US stocks. The inspiration was to create a structured, repeatable research workflow for self-directed investors where a rules-based model scores the data, and AI simply explains it in plain English.
What it does
Qualtix is a model-backed stock research platform built for self-directed investors. It evaluates business quality, valuation, entry timing, risk, and data quality across 1,000+ major US stocks.
- Analyzes individual stocks and generates research briefs in plain English.
- Compares companies side by side.
- Features a 1,000+ stock screener and daily updates.
- Tracks watchlists and provides portfolio analysis.
Instead of acting as an AI stock picker, the platform separates the core metrics into clear model outputs. AI is used only to explain the model's results, making the research process easier to understand without inventing a recommendation.
How I built it
I built the core rules-based model to handle the heavy lifting of financial data processing, risk flagging, and valuation context. Then, I integrated an AI layer specifically constrained to explaining the model outputs rather than generating its own financial advice.
Challenges I ran into
One of the biggest challenges was keeping the AI in check. It is very easy for AI to hallucinate or start acting like a financial advisor. I had to strictly engineer the prompts and backend to ensure the AI only explains the model's scores and never acts as a trading bot or buy/sell signal service.
Accomplishments that I'm proud of
I am incredibly proud of how clean the research workflow is. Users do not have to jump between screeners, charts, financial statements, and spreadsheets anymore. The structured research brief successfully turns model outputs into something anyone can read and understand.
What I learned
I learned a lot about the balance between data and presentation. Providing too much raw data overwhelms people, but providing just an AI answer removes trust. The "model scores, AI explains" approach proved to be the perfect middle ground for a business quality stock screener.
What's next for Qualtix
Next up is expanding the screener capabilities, adding more portfolio context features, and refining the Qualtix Coach to provide even deeper insights into valuation and entry timing.
Qualtix is for research and education only and does not provide personalized financial advice.
Built With
- cursor
- next.js
- postgresql
- vercel
