"And I only am escaped alone to tell thee."— Moby-Dick, Epilogue
Survivorship-bias-free, point-in-time ETF holdings built directly from SEC N-PORT filings — delisted names included, identifiers repaired, schemas normalized. Free open-source MCP server. Full point-in-time history bundles coming soon.
Most holdings data quietly deletes the ships that sank. Every merger, delisting, and blow-up that once sat inside an ETF vanishes from the record — and your backtest inherits the bias.
Built from SEC N-PORT filings — the monthly, audited disclosures every US ETF must file. Not scraped fund pages. Not a vendor's re-sale. The filing itself.
Every snapshot is what the fund actually reported for that period — holdings, weights, shares, market values — with the filing date preserved so you know exactly when the market knew it.
Delisted, acquired, and renamed securities remain in the record with provenance-stamped identifier resolution. Survivorship bias is the product's reason to exist.
An MCP server that gives Claude, and any MCP-capable agent, the complete current holdings of any US-registered ETF plus a full identifier crosswalk. Apache-2.0. Runs locally.
| Tool | What it does | |
|---|---|---|
get_etf_holdings | Complete N-PORT holdings for any US ETF | free |
lookup_cik | Ticker → SEC CIK + entity metadata | free |
resolve_ticker | Raw ticker + exchange → normalized symbol | free |
resolve_cusip | CUSIP → ticker / FIGI | free |
resolve_figi | FIGI → ticker + venue | free |
get_etf_holdings_history | Full point-in-time history — the Pequod database | coming |
Per-ETF point-in-time holdings history for the N-PORT era (2019–present), refreshed quarterly. Parquet + CSV, versioned schema, provenance-stamped identifiers (ISIN · FIGI · ticker). No API required — download and join.
A complete PIT history for one ETF, free — full schema, full history, so you can evaluate the data before buying anything.
Download sample →Single funds and fleet bundles, one-off or subscription. Survivorship-bias-free by construction, SEC-primary-source, schema-versioned so your pipelines never break silently.
Get launch pricing →Launch announcements, sample data, and the story of how this dataset is built — on the AB Evergreen Analytics Substack. Subscribe free and you're on the waitlist.
Subscribe on Substack →Or email info@ABAnalytics.dev