The AI research bench
for serious part-time investors.
Type a thesis — in plain English. AlgoThesis pulls real-time data, searches the web, cross-references FINRA short interest, and maps your view to a set of tickers you can actually trade. No advice. Just sharper research.
20 theses/day free · no credit card · $29/mo for unlimited + Python code export
Post-training-cutoff news, grounded.
Ask about a Fed nominee, a tariff change, a CEO shakeup — AlgoThesis runs a real Google Search and cites sources. ChatGPT makes it up from 2024.
The data ChatGPT doesn't have.
Short percent of float, days-to-cover, last report date — sourced from FINRA via Yahoo for 2,400+ tickers. Filter to real squeeze candidates in seconds.
New IPOs show up the day they list.
CRWV, NBIS, FIG, TEM — names ChatGPT can't find because its training stopped a year ago. Our DB refreshes daily from Alpaca, Finnhub, SEC EDGAR.
Type what you're actually thinking.
No keyboard shortcuts to learn. No filter menus. Just the question.
short squeeze candidates with >20% float shortedReturns live short-interest data sorted desc. 12 names filtered from 2,400 in ~1s.
what stocks benefit from a dovish Fed?Web-searches current rate sentiment, maps to regional banks + homebuilders + REITs + long-duration tech. Cites Bloomberg / Reuters sources.
small-cap biotech oversold below $2B capLive DB filter: sector=Healthcare, industry=Biotechnology, RSI<30, market_cap<$2B. 8 names with YTD + reason per ticker.
kevin warsh basketPulls his April 2026 13F disclosure AND frames a thesis basket for a Warsh-led Fed. You pick which angle.
What you actually get.
Honest comparison. No "10× faster" without saying 10× faster than what.
| Feature | AlgoThesis | ChatGPT | Bloomberg | TradingView |
|---|---|---|---|---|
| Live web research with citations | ✓ | — | ✓ | — |
| Real short-interest data | ✓ | — | ✓ | — |
| Thesis-first basket builder | ✓ | — | — | — |
| Natural-language screener | ✓ | — | — | — |
| Save theses, track them | ✓ | — | ✓ | ✓ |
| Price | $29/mo | $20/mo | $2,000/mo | $15/mo |
Built for the part-time investor who takes it seriously.
You have a brokerage. You read Stratechery or Matt Levine. You know what a 10-K is, you've opened one, and you check your portfolio more than your horoscope. You don't want a beginner's app and you're not paying $2K/month for a Bloomberg terminal.
AlgoThesis is the middle lane: sharp enough to respect your time, cheap enough to actually use, and honest enough to cite sources when it makes a claim.
Questions
Is this investment advice?
No. AlgoThesis is a research tool, not a registered investment advisor. Every basket is an example expression of your own thesis. Consult a licensed advisor before acting on anything you see here.
Where does the data come from?
Live prices from Alpaca. Fundamentals from Finnhub. Filings from SEC EDGAR. Short interest from FINRA via Yahoo. Web search grounded by Google via Gemini. We don't make up numbers.
What's different from ChatGPT?
ChatGPT works from training data that ends sometime in 2024. It doesn't have current short interest, recent IPOs, or this week's Fed news. AlgoThesis pulls live data from real sources and cites them.
How is this different from Bloomberg?
Bloomberg is $2,000/month and built for a trader's desk. AlgoThesis is $29/month and built for someone who reads 10-Ks on the weekend. The data overlap is ~70%; the pricing overlap is 0%.
Can I backtest?
Yes — from the dashboard. Save a basket, click Backtest. We intentionally don't run backtests inside the chat because 'type a thesis, get 3 years of simulated P&L' feels like we're predicting the future. We aren't.
Open the bench.
Free tier is generous. 20 theses a day, real data, no card. Enough to decide if this is for you.
Start with a thesis →Not investment advice. Not an RIA. Consult a licensed advisor before acting on anything.