Data API

openbasket is an aggregator. It republishes nothing of its own: every holding comes from a named public source, and every file and every row says which one, with a link and an as-of date.

The site at https://openbasket.leithresearch.org/ is static. The “API” is a set of JSON files rebuilt by the scheduled build and served by GitHub Pages, so there is no server to overload. All paths are relative to the site root.

api/v1/meta.json

{
  "api_version": 1,
  "built_at": "2026-11-03T12:00:00Z",
  "release_tag": "2026-11",
  "release_url": "https://github.com/omaxim/openbasket/releases/tag/2026-11",
  "full_data_url": "https://github.com/omaxim/openbasket/releases/download/2026-11/openbasket-lookthrough.parquet",
  "n_funds": 15234,
  "n_isins": 31877,
  "sources": [
    {
      "id": "ishares",
      "name": "iShares (BlackRock) fund holdings",
      "url": "https://www.ishares.com/",
      "licence": "Public website data; terms not reviewed",
      "credit": "Holdings data from iShares by BlackRock",
      "n_funds": 1913,
      "as_of_min": "2026-10-30",
      "as_of_max": "2026-11-02"
    }
  ]
}

sources[].id is one of ishares, spdr, tickerinside, nport, security_master. Every source that contributes a row anywhere must be listed here, with its credit line.

api/v1/index.json

Every fund the build covers, as a table, so the search box can load it once (gzip on the wire, about 1 MB):

{
  "fields": ["id", "isin", "name", "ticker", "issuer", "kind", "source", "as_of", "resolved_pct", "n_constituents"],
  "rows": [
    ["IE00B5BMR087", "IE00B5BMR087", "iShares Core S&P 500 UCITS ETF", "CSPX", "iShares", "fund", "ishares", "2026-11-02", 99.86, 503]
  ],
  "aliases": {"IE00B5BMR087": "IE00B5BMR087"}
}

• id names the fund file. It is the fund’s ISIN where one is known, otherwise its SEC series ID (S000004310).

• isin may be null (N-PORT series whose share-class ISINs are unknown).

• aliases maps every known ISIN, including other share classes of the same portfolio, to an id.

• kind is fund for everything in the index; single bonds and stocks are not listed (they are their own constituent).

api/v1/funds/{id}.json

{
  "id": "IE00B5BMR087",
  "isin": "IE00B5BMR087",
  "isins": ["IE00B5BMR087"],
  "name": "iShares Core S&P 500 UCITS ETF",
  "ticker": "CSPX",
  "issuer": "iShares",
  "source": "ishares",
  "source_url": "https://www.ishares.com/uk/individual/en/products/253743/...",
  "credit": "Holdings data from iShares by BlackRock",
  "as_of": "2026-11-02",
  "matched_via": "ishares",
  "note": null,
  "reported_total_pct": 100.00011,
  "resolved_pct": 99.86422,
  "unresolved_pct": 0.13578,
  "unidentified_lines_pct": 0.14602,
  "n_constituents": 503,
  "truncated": false,
  "fields": ["isin", "name", "weight_pct", "asset_class", "source", "as_of", "depth", "via"],
  "constituents": [
    ["US0378331005", "APPLE", 7.59332, "equity", "ishares", "2026-11-02", 0, "iShares Core S&P 500 UCITS ETF"]
  ],
  "unresolved": {"weight_pct": 0.13578, "note": null},
  "full_data_url": "https://github.com/omaxim/openbasket/releases/download/2026-11/openbasket-lookthrough.parquet"
}

• constituents are sorted by absolute weight, largest first. Weights are percent of the fund’s net assets and are never rescaled.

• Each row’s source and as_of can differ from the fund’s (a fund of funds mixes sources); via names the fund the row came from.

• Large funds are cut to the build’s max_constituents rows; truncated is then true and the full list is in full_data_url.

api/v1/funds/{isin}.json for other share classes

A share-class ISIN that is not a fund file’s id gets a small alias file:

{"alias_of": "IE00B5BMR087"}

Release assets

Each scheduled build publishes a GitHub Release tagged YYYY-MM with openbasket-lookthrough.parquet (every constituent row of every fund, no truncation), openbasket-funds.parquet (one row per fund, the summary fields above) and quality-report.json.