Open aryan1107 opened 2 weeks ago
We’re definitely planning to support this feature in the coming months. Just curious, what’s your use case that requires this support?
+1 (also need that... I'm now feeling like an unpaid Livekit developer xd I've been working with the live kit interface for half a year now.)
We’re definitely planning to support this feature in the coming months. Just curious, what’s your use case that requires this support?
This is relevant for guaranteed schemas for function calling as well as COT or moderation use cases. As an example, we have several scenarios we need to flag for every response, so rather than calling a function each time we have flags using structured outputs that we check before the model response.
Wondering if new support for Structured Outputs can be added for openai models?