What
The settings schema is the generated description of which values each setting accepts. It now adds old alias values to the allowed list for inferenceCredentialKind and marks them under x-deprecatedValues, each with a warnFrom date and a supportEnds date.
Why
Old values still count as valid, but the schema now records that they are on the way out and when.
Names in the bundleinferenceCredentialKind
Documented inclaude-docs/cowork/changelog
The entry above is what we published on the day. These lines were added later, as Anthropic's own pages caught up, and they sit beside the original rather than replacing it.
Confirmed since
Anthropic's documentation has since written up inferenceCredentialKind, on Changelog.
* Changed `inferenceGatewayAuthScheme: "sso"` to have no end date in this release and later: if `inferenceCredentialKind` is not set, they keep reading `sso` as `inferenceCredentialKind: "interactive"` after October 7, 2026. Earlier releas…cowork/changelog see the edit
Added since
A small documentation edit on Deploy Claude Desktop on 3P with an LLM gateway touched a line naming inferenceCredentialKind after this was published.
Earlier app versions used `inferenceGatewayAuthScheme: "sso"` to select this mode. That value is deprecated; set `inferenceCredentialKind: "interactive"` instead. Existing deployments that still send `inferenceGatewayAuthScheme: "sso"` wor…third-party/claude-desktop/gateway see the edit
One source agreesOne thing we can check says the same as this entry.
Anthropic's documentation agrees
A small documentation edit on Deploy Claude Desktop on 3P with an LLM gateway touched a line naming inferenceCredentialKind after this was…