stocks-llm helps you discover public companies through natural-language themes, filters, side-by-side comparisons, and company profiles, each backed by dated, cited evidence from SEC EDGAR filings and market-data fundamentals.

Where does stocks-llm get its data?

Company filings and profile facts come from SEC EDGAR; fundamentals and prices come from licensed market data. Figures carry their source and as-of date, and market data is delayed daily-close — not real-time.

Why is a company I searched for not showing up?

The catalog is not exhaustive. If a public company is not in it, that means it is not currently covered — it does not mean the company was excluded, delisted, or fails to qualify. Coverage expands over time as constituents and data are refreshed, so a missing company may simply not be ingested yet.

Is stocks-llm financial advice?

No. stocks-llm is informational only and is not financial advice. Data may be delayed, missing, or stale, and coverage varies. Verify information independently and consult a qualified professional before making any decision.

Does stocks-llm track congressional stock trading?

Yes. stocks-llm ingests U.S. House of Representatives STOCK Act periodic transaction reports (from the Office of the Clerk) and shows, per company, the disclosed Congressional trades — member, transaction type, the disclosed amount range, and the filing date — on each company’s "congress-trades" page. These are lagged legal disclosures (filed up to 30–45 days after the trade), not real-time trades, and are informational only, not a buy or sell signal.

Does stocks-llm track insider trading (SEC Form 3/4/5)?

Yes. stocks-llm ingests SEC Form 3/4/5 filings and shows, per company, the insider (officer, director, or 10% owner) transactions — buys and sells with their filing dates and links to the original SEC filing — on each company’s "insider-trades" page. Insider filings are lagged legal disclosures, not real-time trades, and are informational context, not investment advice.

Can AI agents use stocks-llm?

Yes. stocks-llm exposes a read-only MCP (Model Context Protocol) server over Streamable HTTP at https://stocks-llm.com/mcp with the tools search_companies, get_company_profile, and compare_companies. No authentication is required and it has its own rate limit.

About stocks-llm

Product tour (3:42): plain-English questions, comparisons, technicals, chart patterns and the Markets pages. Figures on screen are as of the 28 Sep 2026 close.

Research stocks in your own words.

stocks-llm helps you discover public companies by describing what you want in plain English — a theme, a filter, or a head-to-head comparison — and get real companies back, each with an evidence-grounded "why" drawn from SEC filings and fundamentals, every figure carrying its source and as-of date.

Data last updated 2026-10-08.

Try a natural-language search → Market movers → Your portfolios →

How it works

Describe what you're looking for — for example "medical companies working on diabetes" or "profitable AI-exposed large caps" — and stocks-llm returns matching real companies from its catalog, each with a short factual rationale and a card (ticker, delayed daily-close price with its as-of date, sector, and links to filings and the company website). You can also browse and filter, open a full company profile, or compare two companies side by side.

Ask — describe a theme in plain English and get matching companies back, each with a dated, sourced rationale.
Top 10 — a scored list of ten companies, rebalanced on a calendar and tracked against the market.
- Top 10 — ten companies ranked by the published score, rebalanced monthly.
- By sector — the same list again, one per sector.
- How the score works — the score's formula and weights, and the rule that moves a company in or out.
Screens — ready-made screens over the catalog: technical ones recomputed nightly, fundamental ones ranked on dated figures.
Patterns — classical chart shapes found in our own stored daily closes and drawn on the price they were found in.
Browse — filter the catalog by sector, index membership, size, valuation, growth and more.
Markets — daily destinations, each built from data the site already stores.
- Movers — the biggest 1-year gainers and losers, and the largest companies.
- Highs & lows — companies printing a new 52-week high or low.
- Market health — how much of the catalog sits above its own moving averages.
- Squeeze watch — high short interest shown next to the price and volume conditions around it.
- Momentum — companies beating and lagging the market, over your choice of window.
- Earnings radar — who reports next, the largest recent surprises, and the beats the price fell on anyway.
- Analyst changes — price-target raises and cuts, each dated to the two snapshots that moved between.
- Dividend calendar — ex-dividend dates, with filed dates and our own estimates labelled apart.
- Investors — what large institutional managers reported holding on their latest SEC 13F, and what changed since the one before.
- ETFs — the US-listed funds we cover and what each one tracks, with its latest daily close.
Blog — daily market notes and longer essays, every figure linked to the record behind it.
Portfolios — save companies for later. Signed in, your named lists follow you to any device; signed out, they are stored in your browser only.

Reached from a company or a result rather than the menu:

Company profiles (e.g. /company/AAPL) — business summary, fundamentals, dividends, technicals and peers, every figure dated and sourced.
Head-to-head comparisons (e.g. /compare/aapl-vs-msft) — two companies’ real metrics side by side.
Collections (e.g. /companies/tech-stocks) — companies grouped by sector, theme and popular search term.
Filing archives (e.g. /company/AAPL/filings) — every filing we hold for a company, linked to the original on SEC EDGAR.
Institutional holdings (e.g. /investors/berkshire) — what a large investment manager reported holding on its latest SEC Form 13F, and what changed since the quarter before.
Insider trades (e.g. /company/AAPL/insider-trades) — SEC Form 4 transactions, classified by code rather than lumped together as buying or selling.
U.S. House disclosures (e.g. /company/NVDA/congress-trades) — trades disclosed by U.S. House members under the STOCK Act, in the amount bands the filings themselves use.
What moved a price (e.g. /company/AAPL/moves) — a dated price move and the filed events that fall in the window ending at it, in time order — candidates, never a single asserted cause.
Similar companies (e.g. /similar/NVDA) — the peers of one company, from the same list the chat answers “companies like X” from.
Signal indexes (e.g. /signals/near-52-week-high) — one page per disclosure signal we compute, each using the same predicate the screener applies.
Indicator reference — every technical indicator this product computes, derived from the registry that defines them.
Glossary — plain-English definitions of the terms the site uses, each linked to where it is used live.
Getting started — what the product does for someone who has not met a financial ratio yet, in runnable examples.
Question-led guides (e.g. /natural-language-stock-screener) — one page per question people actually ask, each answering it with real queries into the product.
Weekly digest — a dated recap of the week’s disclosures, every claim linked to its source.
MCP integration — the endpoint an AI client adds to read this catalog directly, with its tools and connection details.
Shareable results — any answer can be frozen at a permanent /r/ link that keeps the figures and their as-of dates as they were read.

Ready-made screens

82 screens run over the covered catalog — technical ones recomputed nightly from stored daily closes, fundamental ones ranked on dated figures. Selection is deterministic: a company appears only when its own stored data meets the screen's rule, and a company with no data for that metric is excluded rather than estimated. Browse them all at /screens.

Trend (moving averages) (4) — e.g. Trading below the 200-day moving average
Momentum (RSI) (2) — e.g. Oversold (14-day RSI below 30)
Crossovers (2) — e.g. Recent death cross (50-day crossed below 200-day)
52-week range (3) — e.g. Pulled back from the 12-month high
All-time range (2) — e.g. Near the all-time high
Volume (1) — e.g. Unusually heavy trading volume
Analyst estimates (2) — e.g. Analyst price target recently CUT
Valuation & size (3) — e.g. Largest companies by market cap
Profitability (16) — e.g. Highest return-on-invested-capital companies
Growth & performance (3) — e.g. Best 1-year price return
Dividends (2) — e.g. Highest dividend-yield companies
Breakouts (3) — e.g. Broke down to a new 20-day low
Candlestick patterns (4) — e.g. Printed a bullish engulfing pair
Elliott wave (6) — e.g. Elliott wave 3 candidates
Momentum (2) — e.g. MACD just crossed bullish (12/26/9)
Momentum (stochastic) (2) — e.g. At the bottom of its 14-day range (stochastic below 20)
Price gaps (4) — e.g. Left an unfilled gap (5%+ in the last week)
Price streaks (2) — e.g. On a winning streak (5+ higher closes in a row)
Range (Fibonacci retracement) (1) — e.g. Near the golden pocket (has given back about two thirds of its 52-week move)
Range (support & resistance) (2) — e.g. Near a resistance level (pressing into a price that has capped it before)
Relative strength (2) — e.g. lagging the market
Signal agreement (2) — e.g. Most technical signals bearish
Trend (2) — e.g. trendless / choppy (low ADX)
Trend (Ichimoku cloud) (3) — e.g. Trading below the Ichimoku cloud
Volatility (5) — e.g. in a Bollinger Band volatility squeeze
Volume-weighted price (2) — e.g. Under water since its last earnings report (below the anchored VWAP)

You can also just ask for one in plain English:

How the numbers work

Every number on stocks-llm is grounded in real data and timestamped — figures are never invented. Company filings and XBRL financial facts come from SEC EDGAR; fundamentals and prices come from dated market data. Prices are delayed daily-close values, labeled with their as-of date — stocks-llm does not provide real-time or intraday prices. Where a figure is missing it is shown as "n/a", never estimated.

Company descriptions and the natural-language matching are generated with an LLM over that real underlying data; recommendations are limited to real companies in the catalog with real cited figures. The model never invents a ticker, a price, or a figure: every number shown is read from the stored record and carries its own as-of date, and a company the catalog does not cover is reported as not covered rather than described from memory.

What's covered

The catalog covers public companies in the markets stocks-llm offers: United States — S&P 500 + NASDAQ-100 + Dow + Russell 1000 + Russell 2000; United Kingdom (LSE) — FTSE 100 + FTSE 250. Approximately 3,197 US and 251 UK companies in the current catalog as of 2026-10-08, deduplicated by ticker. This is a snapshot; the count changes as constituents and data are refreshed.

Coverage and enrichment vary by company — some figures may be missing where a source does not report them (shown as "n/a", never guessed).

The catalog is not exhaustive. If a company you look for is not in it, that means it is not currently covered — not that it was excluded, delisted, or fails to qualify. Coverage expands as constituents and data are refreshed.

For AI agents: MCP server

stocks-llm exposes a read-only MCP (Model Context Protocol) server over the Streamable HTTP transport so other people's AI agents (Claude Desktop, Claude Code, or any MCP-compatible client) can query the same real, dated, cited data directly. No authentication is required (it serves the same public data already available over the web), and it has its own rate limit. Three read-only tools are exposed:

search_companies — natural-language / filtered company search.
get_company_profile — the full profile for one ticker.
compare_companies — side-by-side metrics for two tickers.

Add it to an MCP client with the endpoint URL alone:

https://stocks-llm.com/mcp

See /llms.txt for a machine-readable map of the site and its key surfaces.

Insider & U.S. House stock-trading disclosures

stocks-llm surfaces two kinds of legally-required public trading disclosures, per company, sourced directly from official filings:

Insider trading (SEC Form 3/4/5) — buys and sells by a company's officers, directors, and 10% owners, from SEC EDGAR, shown with filing dates and links to the original filing on each company's insider-trades page (e.g. /company/AAPL/insider-trades).
Congressional trading (STOCK Act) — U.S. House of Representatives periodic transaction reports from the Office of the Clerk, showing the member, the transaction type, the disclosed amount range, and the filing date on each company's congress-trades page (e.g. /company/AAPL/congress-trades).

Both are lagged legal disclosures — filed days to weeks after the trade, not real-time — and are shown as informational context only, never as a buy or sell signal. U.S. House disclosures report broad amount ranges (as legally required) and may cover a member, spouse, or dependent.

Limitations & disclaimer

Informational only — NOT financial advice. stocks-llm is a discovery and research aid, not a recommendation to buy or sell anything. Market data is delayed daily-close, not real-time. Figures may be missing, stale, or incorrect, and company coverage and enrichment vary. Nothing here is a solicitation or an offer. Always verify information independently against primary sources (SEC EDGAR) and consult a qualified financial professional before making any investment decision.

Home · Movers · Tech stocks · AAPL vs MSFT · Portfolios · Glossary · llms.txt