namespace for compatibility; the playground labels it Dataset.
All three spellings remain valid for REST reads. An already configured custom namespace takes precedence over a short alias of the same name. Account entitlements retain their established namespace values. Aliases grant no features, sources, regions or credits.
pilot is a dataset name, not a promise of in-play or current coverage. Stream subscriptions accept the documented dataset aliases while retaining stored authority and exact native identity. Management bodies keep their documented namespace spellings.
Consistent fixture and odds responses
Existing calls retain their original response shape by default. Addview=standard to fixture or odds reads for common DatasetFixture and DatasetOdds records. Explicit view=legacy keeps the original format. Both use the existing data, missing[] and meta envelope, paging, authentication and credit rules.
sportsbooks dataset, select exactly one reviewed source using source_id. Sources5/14/15 use the Kambi reader and require kambi_native_retained; live mode additionally requires the existing founder live approval and supports14/15 only. Sources11/12/13 require native_odds. These grants remain separate from canonical fixtures and odds grants. Unsupported native filters fail instead of silently switching readers. The old /v1/native/... routes also accept the standard view.
Fields and identity
- Fixtures expose
id,identity_scope, optional source/sport/competition IDs, schedule, status, participants and evidence. Unknown metadata stays null or empty;originalretains the existing full record and bindings. - Odds expose fixture/source/market/selection IDs, names, class, period, line/unit, side, value, state, participant evidence and verification flags. Stale or suspended prices are not promoted into the common active
valuefield. The original record retains its explicit historical/diagnostic state. - Resource identifiers in the standard view are exact strings, including leading zeros. Registered source IDs and known canonical sport/competition IDs retain their numeric types. Legacy responses keep their original ID types.
- Compare a line together with
line_unit: canonicalquarter_pointsare integer quarter units; native lines retain their documented unit. Equal-looking numbers are not a settlement compatibility decision. evidencepreserves capture, receive, confirmation and source times where available. Retained evidence does not become current when delivered or replayed.meta.datasetgives the preferred short dataset name for standard responses;meta.namespaceretains its established representation.
Participants
kind=team is the default. Canonical participants are restricted to accessible collected fixtures and associated permitted markets; unrelated catalogue rows are excluded. Native teams come from admitted fixture evidence. Native player reads require a fixture and, for Kambi, its exact capture. Missing player identity returns an explicit missing reason rather than a fabricated ID.
Participant IDs can be null for name-only native evidence. Non-null native IDs keep their provider scope, source and fixture context, and identity_verified remains explicit. Names and native IDs are never merged across books. These are bounded observations, not a complete roster or league-wide player catalogue.
Fixture detail
fixture_ambiguous; supply the exact capture_id instead of selecting an arbitrary historical snapshot. A detail read needs the same underlying feature/source scope as fixture discovery. It does not return scores, settlements or odds, and does not grant those capabilities.