I'm a data scientist and amateur investor. I've been using OpenAI's Deep Research to get quick overviews of investment candidates, and I saw an opportunity for a version specialized for stocks.
So in the last month I built an MVP with a simple UI: pick a stock (or multiple!), click "research," and in ~5 minutes it produces a report. Here's the demo: https://www.youtube.com/watch?v=2wCM44B6iH8
At first I just wanted to build a tool that automatically integrates SEC filings with OpenAI's Deep Research API to save time on manual uploads. But their API is insanely expensive (up to $10 for one run!), so I decided to build my own agents. They currently do the following:
- Parse data from SEC (10-Ks and 10-Qs)
- Search the web for relevant data (I made a list of industry-specific publications for each sector — agents only search those links and avoid market news like CNBC)
- Pull up-to-date financial data from Financial Modeling Prep
The final combiner agent takes data from these sources and produces a structured report with three parts:
- One-pager for a quick glance
- Full report with citations for a deep dive
- Financials section with graphs to see trends
I also built a dashboard where you can organize reports in folders and easily compare them.
This is NOT a stock tip generator or a replacement for your own judgment. Think of it as a scalable junior analyst who does the research part for you.
I'd love feedback on how this fits into your research workflow. Since API costs are still high, I require a signup, but you get 3 free research credits to try it out.
If you don't feel like signing up, you can comment a ticker you'd like to see and I will share the report in a reply.
