Data & schema semantics · Naming & aliasing

x-obmcs-endpoint-service-name

`x-obmcs-endpoint-service-name` appears at the root of the document. Its value is a string. Used by a single provider across 7 OpenAPI documents. It sits in the `x-obmcs-` namespace, so its meaning is defined by obmcs tooling rather than by OpenAPI. Observed values include `ocir`, `filestorage`, `inference.generativeai`, `generativeai`. Consistent with naming & aliasing — this is inferred from where the key appears and what it carries, not from any published definition.

Data & schema semantics x-obmcs- namespace derived description
How this description was produced. It is assembled from what was measured in the corpus — where this key appears in a document, what shape its value takes, the values observed, and how many providers use it. It is not taken from a published definition, because for most extensions none exists. Read it as evidence, not as a specification. If you own this extension and want it described properly, tell us.
7 occurrences
7 documents
1 providers
vendor-named

Where it appears

Location in the documentOccurrences
root 7

What its value looks like

Value shapeOccurrences
string7

Observed values

Sampled from the specifications, most frequent first.

ocirfilestorageinference.generativeaigenerativeaicontrol-center-dscontrol-center-cppsasvc

Providers publishing it

All 1.

oracle

Why this is not in OpenAPI

Extensions exist because a provider needed something the specification would not carry. Most of what they hold is not the API contract at all — it is operational metadata about the contract: documentation, lifecycle, policy, provenance, and now agents. That metadata is usually better placed alongside the contract, in an Overlay or an APIs.json index, than crammed inside it. See everything else doing the data & schema semantics job.