Skip to content
QuantGPT
← All datasets

US stocks

Security master

QuantGPT

Every US-listed name with its exchange, industry, listing and delisting dates and type flags.

What it is

Every US-listed name with its exchange, industry, listing and delisting dates and type flags.

  • securities: The security master: name, FIGI, CUSIP, exchange, GICS levels, IPO and delisting dates, active, ADR, country.
  • security flags: Type flags per ticker: warrant, unit, preferred, note, ETF, CEF, SPAC, royalty trust, test, non-operating, the vendor's type.
  • gics classification: GICS sector, industry and sub-industry and domicile per ticker.
  • security parent: Share classes mapped to one issuer.
  • security ids: Security ids matched to tickers with the evidence.

Source and licence

Source: EODHD reference data (GICS is an MSCI and S&P scheme).

Licence: licensed vendor: reference fields. This is our reading of the table; counsel decides.

Credit: each company's own 10-K or 20-F on SEC EDGAR, Wikipedia short descriptions and articles (CC BY-SA) through the MediaWiki API, Wikidata descriptions (CC0); industries: GICS, via the QuantGPT engine (Data sources).

Coverage and gap years

46,973 names, 46,973 rows in securities.

Years with no rows: none between the first and the last.

TableRowsFirstLast
securities46,973--
security flags46,973--
gics classification31,970--
security parent3,153Oct 9, 2026Oct 9, 2026
security ids3,994Mar 19, 1998Oct 7, 2026

Freshness

Its rows carry no date. Updates: as listings change.

Point in time

One row per ticker as it stands today, with its IPO and delisting dates.

Not kept: a row is updated in place (it carries created and updated times).

Delisted names

Included: a delisted name keeps its row, with its delisting date and an inactive flag.

Known issues

  • security flags holds 46,973 names. Flags: Rights have no flag of their own. 53 active names called "Right" are flagged neither warrant nor unit. non operating does catch the two seen in the gainers below. (ACCURACY, 6. The security list)
  • Exchange labels are not normalised: "Nasdaq" (18,559) and "NASDAQ" (3,322, of which 304 are active). "NYSE Arca" and "NYSE ARCA". "NYSE American", "NYSE MKT" and "AMEX". "Cboe BZX" and "BATS". The active "NASDAQ" names include OTC-style names, such as five-letter F tickers like RITRF, so a filter on exchange NOT LIKE '%OTC%' lets them through. (ACCURACY, 6. The security list)
  • The 30 biggest one-day gains on 2026-10-08 among active, non-OTC names include: four warrants (FTRA.WS +1,874%, SKYH.WT, SLND.WT, NE.WTA); two rights (EURKR, RFAIR); three leveraged single-stock ETFs (AAOZ, COHQ, AXTQ); the unadjusted MIN split; sub-penny names labelled NASDAQ: TAAG 0.009, RITRF 0.012, PHSE 0.0127, MEHA 0.0054. (ACCURACY, 6. The security list)
  • Four tickers have month-end prices and no securities row. (ACCURACY, 6. The security list)
  • The universe. It comes from listedNames (src/lib/data/engine-market.ts:19,41): security profile joined to securities, active and not OTC, with no warrant, unit, preferred or operating-company test. Nothing in src/ reads security flags or security profile is operating company. (ACCURACY, b. The Screener's Gainers are full of junk warrants)
  • Rights have no flag: 53 active names called "Right" are neither warrant nor unit. (GAPS, 1. Junk-free lists (score 5.0))
  • Exchange labels are not normalised: "NASDAQ" holds OTC-style names. (GAPS, 1. Junk-free lists (score 5.0))
  • 48 unadjusted splits make fake +500% days (ACCURACY.md sections 5 and 6). (GAPS, 1. Junk-free lists (score 5.0))
  • Another agent owns the Gainers issue; this is its data map. (GAPS, 1. Junk-free lists (score 5.0))

Where it is used

Engine tools that reach it (your AI, over MCP):

  • run_screen The CURRENT tickers matching a set of criteria against the latest data in the warehouse - the operational output of a backtest you validated.
  • query_warehouse Run read-only SQL (DuckDB dialect) directly against any research database.

Ask AI

Every count and date here was measured on Oct 9, 2026 from the QuantGPT engine's warehouse and is read from the catalogue file when the site is built. Metadata only: no vendor rows are shown.