Stock Meadow · Guide
AI stock search: find stocks by asking in plain English
Type the stocks or ETFs you're looking for the way you'd say it, and AI search turns the question into filters on Stock Meadow's data for the 3,000 largest US stocks and the largest US-listed ETFs. It shows how it read your question, lists the matching companies with their figures, and lets you refine the list with a follow-up, add it to a list, or score it with a model.
Updated · By Stock Meadow · About us
How AI stock search works
- You ask. Type a question in plain English and press Enter. Simple questions (“utilities”, “largest companies”) are read by code with no AI at all.
- The AI reads it into filters. Sector, sub-sector, company size, metric limits, a theme, a sort order. It never looks up figures and never writes any.
- You see the reading. “We read this as” shows each filter as a chip. Press × on one to drop it. Anything we have no data for is listed under “Couldn't use”, never dropped silently.
- Stock Meadow's data answers. The filters run on the same data as the stock screener, and every stock that matches is listed with the figures you asked about as columns.
Example questions
- “Profitable software companies growing revenue over 20% with little debt”
- “Uranium miners”
- “Companies like COST but smaller”
- “Stocks Warren Buffett owns”
- “Low-fee bond ETFs”
- “Who are the most active traders in Congress?”
- “What are the interest rates?”
- “How do I backtest a model?”
Funds and people work too: “Vanguard international ETFs”, “ETFs holding NVDA over 5%”, “which Super Investors own Apple”, “the most active insiders this month”.
Other kinds that work: “Stocks Warren Buffett owns”, “software companies with revenue over $1B”, “cheaper than its sector, top 10% for ROIC”, “what is Target's revenue and EBITDA”, “stocks in ARKK”, “companies reporting earnings this week”.
What you can ask about
| Kind of question | What it uses |
|---|---|
| Metrics | Any of the 103 metrics in the screeners: P/E, ROIC, margins, growth, debt, dividend yield, momentum and technical indicators |
| Dollar figures | Revenue, EBITDA, net income, free cash flow, debt and cash from the latest statements (US-dollar reporters) |
| Themes | Words matched in the companies' own descriptions, plus short tags an AI wrote for each company, labelled “AI tag” |
| “Like” a company | Its similar companies and sub-sector, ranked by how close their size, growth, margins, valuation and price trend are |
| Compared with others | “Cheaper than its sector”, “top 10% for ROIC”, worked out from today's figures |
| Our other data | Congress and insider trades, Super Investors' 13F holdings, the earnings and dividend calendars, ETF holdings |
| ETFs | Ask for ETFs or funds: kind (bond, international, sector, commodity), fund company, fee, assets, yield, returns, risk, the funds holding a stock, themes in fund names and AI tags. The same figures as the ETF screener, which a search can open with the same filters |
| People | The most active members of Congress (all, senators or representatives, or for one stock), the most active insiders and their largest purchases and sales, the Super Investors who own a stock or have the largest portfolios: a ranking from the trackers, with a link to the full list |
| Stock Meadow itself | “How do I backtest a model?”, “what are the interest rates?”, “where is Ro Khanna's profile?”: answered with links to the pages, tools and user guide sections that answer it. Every other answer also lists its related pages |
| Scoring models | “Score them with GARP” ranks the results with one of the 9 scoring models and adds a Score column |
Swipe sideways to see every column.
Refine the results with follow-ups
After a search, keep typing to change the list rather than starting again: “only under $50B”, “add dividend yield”, “sort by ROIC”, “remove the PEG column”. Choose New search to start fresh. The questions stay in the page address, so reloading and Back keep your place. Columns can also be added, removed and sorted by hand, without using AI.
Where the figures come from
- The AI only reads your question. Every stock and figure shown comes from Stock Meadow's data, the same data as the stock pages and screeners, with its date.
- Check the reading. A question can be misread. The chips show exactly what was searched, so a wrong filter is easy to spot and remove.
- It keeps learning. Each week, questions that found nothing are read again more carefully, and a clearly better reading is kept for the next person who asks.
- Your questions. Questions are stored for up to 90 days without your name or email, to improve the readings. See the privacy policy.
What to do with the results
- Add to list: tick stocks and save them to one of your lists, for the portfolio builder or the scoring model.
- Open in the stock screener with the same filters, to fine-tune them by hand.
- Score them in the stock scoring model, then backtest the model on 20 years of US stock history. Backtest results are simulated; past results don't predict future results.
For research only, not investment advice.
Frequently asked questions
- What is an AI stock screener?
- A stock screener you talk to. Instead of picking filters from menus, you type what you're looking for, such as "mid cap industrials with ROE above 15%", and the AI turns the words into the same filters a classic screener uses. In Stock Meadow the AI only reads the question; the list of stocks and every figure in it come from Stock Meadow's own data.
- Can the AI make up numbers?
- No figure on the results page is written by the AI. It fills in a fixed form (sector, size, metric limits, sort order), that form is checked against the metrics and sectors we have, and the stocks and figures are then read from our database. What the AI can do is misread a question, which is why the filters it read are shown as chips you can check and remove.
- What kinds of questions can I ask?
- Any of the metrics in the screeners (valuation, profitability, growth, debt, dividends, price trend), dollar figures from the statements, themes such as "uranium miners", "companies like Costco but smaller", one company's figures ("what is Lululemon's P/E"), and our other data: Congress and insider trades, Super Investors' holdings, earnings and dividend dates, and ETF holdings.
- Is AI stock search free?
- It is part of the membership, with a 7-day free trial and no card needed to start. Members have a daily limit on searches the AI reads; simple questions and questions asked before are read without AI and don't count towards it.
- Can AI search find ETFs?
- Yes. Ask for ETFs or funds and it searches the 2,000 largest US-listed ETFs with the ETF screener's own figures: "low-fee bond ETFs", "Vanguard international ETFs", "ETFs holding NVDA over 5%", "uranium ETFs". Leveraged and inverse funds are left out unless you ask for them. Without the word ETF or fund, it searches stocks.
- Can I ask about members of Congress, insiders or Super Investors?
- Yes: "who is the most active trader in Congress", "which senators traded NVDA", "the biggest insider purchases this month", "which Super Investors own Apple". The answer is a short ranking from the same figures as the Congress, insider and Super Investor trackers, with a link to the full list.
- Does AI search give investment advice?
- No. It finds the stocks that match what you asked for and shows their figures, so you can research them. For research only, not investment advice.
Ask your first question
Sign up free for a 7-day trial, no card needed, and open AI search from the Research menu. Already a member? Go straight in.
Open AI search