Is it just me, or is the latency for calling MCP servers (e.g. for a new ChatGPT Apps SDK App) quite long? I’m seeing it take > 5s between a pretty direct request (a chat query in the chatgpt app) to call my tool and seeing any request hit my server, as well as delays later before data from useOpenAIGlobal is actually populated into the widget. It looks like a platform behavior, but I wanted to check:
- if this is normal
- if there’s reason to think it will change before broader apps sdk launch
As-is, having this kind of delay really makes it tricky to build for great user experiences.