Get financial facts

Facts endpoints return the verified numbers. There are three views over the same data, each in whole-company form and in by-series form:

Endpoint	Returns	Use when
`POST /api/b2b/facts/core`	Every consolidated fact (no segment/product slices)	Building statement-style views
`POST /api/b2b/facts/latest`	The most recent fact per consolidated series	“Current snapshot” cards
`POST /api/b2b/facts/full`	Everything, including dimensional slices	Segment/geo/product analysis
`…/by-series` variants	Same shapes, restricted to specific `series_key`s	You already searched and know what you want

All are POST with a JSON body identifying one company by ticker or cik, plus an optional metric_id filter.

Fetch all consolidated revenue facts

curl -s https://api.stockalloy.com/api/b2b/facts/core \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $STOCKALLOY_API_KEY" \
  -d '{"ticker": "DECK", "metric_id": "us-gaap#RevenueFromContractWithCustomerExcludingAssessedTax"}'

Each row looks like:

{
  "fact_key": "3f6c1f0a9d…",
  "series_key": "91b2c4d6e8…",
  "cik": "0001011570",
  "current_ticker": "DECK",
  "metric_id": "us-gaap#RevenueFromContractWithCustomerExcludingAssessedTax",
  "metric_label": "Revenue",
  "period_key": "duration:2023-04-01..2024-03-31",
  "period_kind": "duration",
  "period_start": "2023-04-01",
  "period_end": "2024-03-31",
  "period_end_year": 2024,
  "period_end_quarter": 1,
  "unit_signature": "monetary:USD",
  "unit_class": "monetary",
  "currency": "USD",
  "value_num": "4287763000",
  "value_scaled_int": "4287763",
  "scale": 3,
  "source_accession": "0001011570-24-000039",
  "source_form_type": "10-K",
  "source_is_amendment": false,
  "source_is_current_filing_period": true,
  "source_filed_at": "2024-05-24",
  "evidence_key": "01HYX…",
  "evidence_url": "https://stockalloy.com/api/evidence/01HYX…/open?cik=0001011570&accession=0001011570-24-000039"
}

Rows are sorted newest period first, then newest filing first.

Exact values, no floats

Values are transmitted as strings so nothing is lost to floating point:

value_num — the exact decimal value in full units (dollars, not thousands), as a string.
value_scaled_int + scale — the same number as an integer significand and a power-of-ten exponent: value_num = value_scaled_int × 10^scale.

In the example above: 4287763 × 10³ = 4,287,763,000 — Deckers’ FY2024 revenue of $4.29B, exactly as filed.

Period semantics

period_kind: "duration" — flows (revenue, net income). period_key is duration:START..END.
period_kind: "instant" — balances (assets, cash). period_key is instant:DATE and period_start is null.
period_end_year / period_end_quarter are calendar year and quarter of the period end — for companies with offset fiscal years (like Deckers’ March year-end), the calendar year of the period end is not necessarily the fiscal-year label the company uses.
To classify a duration row as annual vs. quarterly, use the day span: ~355–380 days is a fiscal year, ~84–99 days is a quarter.

Revenue/income series that mix annual and quarterly filings are normalized server-side so a Q4 slot never silently contains a full-year value.

Provenance fields

Every row tells you exactly which filing “won” for that period:

source_accession — the SEC accession number.
source_form_type / source_is_amendment — e.g. 10-K, 10-K/A. When a company amends or restates, the winner logic serves the authoritative value.
source_is_current_filing_period — true when the period is the filing’s own reporting period, false when it’s a restated prior-year comparative.
evidence_url — the proof link.

Fetching by series key

When you already know the series (from search or the catalog), skip the full export:

curl -s https://api.stockalloy.com/api/b2b/facts/full/by-series \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $STOCKALLOY_API_KEY" \
  -d '{
    "ticker": "DECK",
    "series_keys": ["91b2c4d6e8f0a2b4c6d8e0f2a4b6c8d0e2f4a6b8c0d2e4f6a8b0c2d4e6f8a0b2"]
  }'

Up to 100 series_keys per call; results come back grouped in your requested key order, newest period first within each series. Dimensional rows include dimension_path_label (e.g. Direct-to-Consumer) so you can label charts without touching raw XBRL axis names.

Discovering what exists: the series catalog

POST /api/b2b/series/catalog lists every series a company has, with period coverage and observation counts — useful for building pickers or auditing coverage:

curl -s https://api.stockalloy.com/api/b2b/series/catalog \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $STOCKALLOY_API_KEY" \
  -d '{"ticker": "RKLB"}'

And POST /api/b2b/dims/catalog lists the dimension combinations (axis/member paths) the company reports, with semantic flags (contains_geo_like, contains_business_like, contains_product_like) so you can group slices without parsing XBRL yourself.

One caveat worth knowing: XBRL tag succession

Around 2018 most US companies switched revenue tags (SalesRevenueNet → RevenueFromContractWithCustomerExcludingAssessedTax, the ASC 606 adoption). Raw series are keyed by XBRL concept, so a single concept’s series can show a gap near 2018 — the older data lives under the predecessor tag’s series. If you need a continuous headline-metric history, query both concepts, or use the MCP server’s search_facts with a plain-English metric name, which bridges the succession automatically.