LikeFolio Data
LikeFolio Data · API docs

Bulk downloads ZIP of CSV

The whole dataset in one authenticated download: every company's X series, every trend, or every published search brand. Each drop holds exactly the rows the API serves your tier, with the same ids, columns and clipping.

GET /v1/bulk

Lists the drops for your tier, newest first, at most three per dataset.

ParameterNotes
datasetOptional: company_demand (X company series), trends (consumer trends) or search (Google search interest). Only published datasets are listed.
formatAs usual.

Rows: dataset, version, bytes, sha256 (of the ZIP) and url, the download link for /v1/bulk/download.

curl -s -H "Authorization: Bearer $LIKEFOLIO_API_KEY" "https://data.likefolio.com/v1/bulk"
{ "dataset": "bulk", "rows": 3, …,
  "data": [
    { "dataset": "company_demand", "version": "<release>_full", "bytes": …, "sha256": "…",
      "url": "https://data.likefolio.com/v1/bulk/download?key=bulk/subscription/company_demand/<release>_full.zip" },
    { "dataset": "trends", "version": "<through>_full", … },
    { "dataset": "search", "version": "2026-10-31", "bytes": 26170057, "sha256": "…",
      "url": "https://data.likefolio.com/v1/bulk/download?key=bulk/subscription/search/2026-10-31.zip" }
  ] }

Illustrative. Use each url exactly as listed.

GET /v1/bulk/download

Downloads one listed drop. Send your API key, as on every route; there are no presigned links.

ParameterNotes
keyRequired. The key from a listed url: bulk/<tier>/<dataset>/<version>.zip.
import hashlib, io, os, zipfile, requests, pandas as pd

BASE = "https://data.likefolio.com/v1"
H = {"Authorization": f"Bearer {os.environ['LIKEFOLIO_API_KEY']}"}

drop = requests.get(f"{BASE}/bulk", headers=H, params={"dataset": "company_demand"}, timeout=60).json()["data"][0]
r = requests.get(drop["url"], headers=H, timeout=600)
r.raise_for_status()
assert hashlib.sha256(r.content).hexdigest() == drop["sha256"], "checksum mismatch: refresh the bulk listing and download again"

z = zipfile.ZipFile(io.BytesIO(r.content))
coverage = pd.read_csv(z.open("coverage.csv"))
series = pd.read_csv(z.open("company_demand_sample.csv"))

What is in each drop

company_demand

FileColumnsRows
coverage.csventity_id, ticker, company_id, perm_id, figi, company, division, is_primary, valid_from, valid_toThe default /v1/universe rows: the live roster of core entities plus the delisted and acquired names.
company_demand_sample.csventity_id, ticker, company_id, date, metric, kdate, raw, corrected, flagThe default company-metrics rows (default corrections) of every company the API serves, including the history of companies no longer covered, ordered by ticker and company_id. kdate is the knowledge date of the rows.
manifest.jsonSee below.
FileColumns
trends_sample.csvtrend_id, name, date, raw, corrected, flag: every catalog trend's series, in catalog order.
trend_catalog.csvtrend_id, name, group, mapped_tickers, mapped_datasets
trend_ticker_map.csvtrend_id, name, ticker, weight, added_at, removed_at
trend_dataset_map.csvtrend_id, name, series, sign, primary, freq, desc
manifest.jsonSee below.

The company and trend files have the same names and columns as the published trial sample, so code written against the sample reads a full drop unchanged.

FileColumns
search/<TICKER>.csvOne file per published company: its verified brands' daily rows, as /v1/company/{ticker}/search serves them. Trial drops carry brand_id, not the brand name.
search_coverage.csvticker, coverage, coverage_full_history, verified_brands, candidate_brands: one row per company.

The search drop has no manifest.json; check it against the sha256 listed by /v1/bulk. It is rebuilt with each search export and named by that export's date. Multiple exports on one date can replace the bytes at that same URL. This name is not an immutable version or an archive of prior search snapshots; save both the downloaded bytes and their checksum if you need a reproducible local copy. Bulk ZIPs have no kdate; dated Google search reads are on four API routes.

manifest.json and versions

The company and trends drops carry a manifest.json (the search drop does not): dataset, version, tier, built_at, embargo_cutoff, release_id, kdate (trends: through), published_at, projection, and files with each member's rows, bytes and sha256.

Search bulk/API parity. The search ZIP applies the same client-visible date, ownership, completeness and tier embargo filters as the search API. Compare the same export and tier; a rebuild may replace the date-named ZIP. Matching rows establish bulk/API parity, not independent proof of company identity, source accuracy or point-in-time history.

Current availability and historical access. The current search ZIP uses the same generation and current-company guard as the API. It omits HELD company CSVs and includes availability.json describing ready and held companies. Expected holds do not abort other companies; technical corruption still fails the generation. Legacy history is accessed through the explicit history route, not included as a fallback in the current ZIP. Native-weekly source archives are not served.

Availability endpoint clocks. In each tier’s availability.json, complete_through matches the canonical client API endpoint after identity and tier caps. source_complete_through preserves the underlying policy endpoint. endpoint_scope states canonical client identity and tier cap; historical baskets not certified. These fields do not turn source as_of into global completeness or certify old baskets. Trial embargo remains separate from the 14-day source freshness assessment.

Subscription display metadata. The subscription search ZIP may include display_names.json, separate from canonical CSVs. It contains schema=search-display-names-v1, semantics and brands, keyed by stable brand ID; each entry has ticker, canonical_brand and display_name. Trial ZIPs omit this member. No source queries or approval receipts are exposed. Canonical CSV labels, columns and IDs are unchanged.

X company metric holds. For a company-demand ZIP built from an affected publication, manifest.json adds metric_coverage_issues (objects with ticker, metrics and reason) and complete_tail_scope="unheld_metrics". This differs from the company API issue object, which uses id. Held CSV cells are blank. Older release ZIPs retain earlier values and do not acquire a retrospective hold annotation.

GT search ZIP checksums and safe resume

A GT search ZIP download uses a strong checksum ETag, ETag: "sha256-<ZIP SHA-256>". Retain the quoted ETag with any partial download. Date-named ZIPs can be replaced; a matching name or modification date does not prove identical bytes.

For resume, send Range: bytes=<saved-byte-count>- with If-Range set to that exact strong ETag. When If-Range is supplied, only a matching strong validator permits a valid satisfiable range to return 206 Partial Content. A stale ETag, weak validator (W/), or date-based If-Range returns the full file with 200 OK. Do not append a 200 body to old partial bytes: replace the partial file with the full response.

Append a 206 body only after checking its ETag and Content-Range match the saved file and expected offset. For an unsatisfiable range (416), refresh the listing and restart. Always verify the assembled ZIP against the matching listed sha256; if the publication changed, refresh the listing and download again. These checks establish byte integrity, not immutable hosted versions, complete source coverage or final launch clearance.