StockAnalysis.com Stock & ETF Screener

The entire US stock and/or ETF market from stockanalysis.com's own screener pages, in one or two HTTP requests — no pagination, no per-symbol lookups, no authentication.

---

What you get

market: "stocks" — one SCREENER_SUMMARY + one STOCK row per ticker: symbol, name, market cap, price, day change %, industry, volume, PE ratio. 5,610 tickers as of 2026-08-16.

market: "etfs" — one SCREENER_SUMMARY + one ETF row per fund: symbol, name, asset class, AUM, price, day change %, volume, holdings count. 5,582 funds as of 2026-08-16.

market: "both" (default) — both of the above in one run.

Rows are pre-sorted by upstream in market-cap / AUM descending order.

---

Input

{
  "market": "both",
  "symbols": [],
  "limit": null
}

• market — "both" (default), "stocks", or "etfs".
• symbols — optional ticker whitelist (e.g. ["AAPL", "VOO"]), case-insensitive. Applied after the fetch — upstream doesn't support server-side filtering (verified: query strings don't change the response), so this controls rows returned, not requests made. A symbol not found in the fetched universe gets its own ERROR row.
• limit — optional cap on rows returned per market, applied after any symbols filter. Since upstream is pre-sorted by size, this is an honest "top N by market cap / AUM", not an arbitrary truncation.

Leave symbols empty and limit unset to get the full market (the default).

---

Known limits

1. Column set is fixed. The site's marketing mentions "313 different metrics," but those are only reachable through the site's own client-side JS after the page loads — every query-string combination tested during recon returned an identical server-rendered payload. This actor exposes the 8 fields per market that the server-rendered page actually carries; it does not (and cannot, HTTP-only) request a custom column set.

2. No per-symbol detail pages (financials, dividends, history) yet. This version is the bulk screener only. See CRAWLING_METHOD.md §7 for why, and what a follow-up would look like.

3. change is upstream's raw percentage number (e.g. -0.06 means -0.06%, not -6%), passed through as-is.

---

Errors are honest, no disguised-200 traps

A bad symbols entry doesn't get silently dropped — it produces an ERROR row naming the missing ticker and the size of the universe that was actually searched. SCREENER_SUMMARY rows carry honestCountMatch: upstream's own claimed total compared against what this actor actually extracted, checked fresh on every run (not just at build time), so a future silent truncation on stockanalysis.com's side would show up as false here rather than quietly under-counting.

Anti-bot posture

None encountered. 3/3 curl_cffi TLS profiles (chrome124/safari17_0/ firefox133) clean 200 on both screener paths, byte-identical response sizes across profiles — no Cloudflare/Akamai/DataDome/PerimeterX marker in the response body.

Policy

https://stockanalysis.com/robots.txt has a single User-agent: * group — no ClaudeBot/anthropic-ai/CCBot/GPTBot-specific group at all. Only /e/ (analytics beacon, unused here) and /p/ are disallowed, plus three unrelated named scraper bots blocked outright.