Get the caller's effective model access policy
What the caller may use: the platform, team and own layers of the model access policy combined (each list is the intersection of the layers). data_region is the effective Neo model inference region; it is absent when the layers’ regions do not overlap.
Authorizations
JWT authentication token
Response
Effective model access policy
The caller's effective model access policy (ModelAccessPolicy) and the restrictions that apply to it.
1 True when the policy limits which models the caller may use (a model source, vendor, model, custom endpoint or share list, or a region other than global). personal_ai_connections and default_model do not count. Clients read this instead of deriving it from the other keys.
The refusal reasons that apply to the caller, each with the layer that set it. Empty when nothing is restricted.
Team AI connection shares, by connection ID.
The Neo model inference region; global when absent.
global, us, eu Team layer only. The team's default model.
neo, byok, ai_connections Exact model IDs allowed per model source: "neo", "personal", "share:" or a bring-your-own-key credential ref.
Team layer only. false keeps members off their personal AI connections.
Model vendors, lower case ("anthropic", "openai", "other").

