Open heymartinadams opened 5 months ago
By any chance, would you know how to handle incoming tool outputs on the client, in this streaming scenario?
The assistant runs a tool (e.g. “weather”) and streams a system response to the client/browser (not as part of the text stream) that indicates the need for a particular UI (e.g. a clarification about what location the user is asking about the weather for).
Hi Stephen,
I just want to thank you for posting this OpenAI assistant streaming starter kit 🙏🏼
It helped me a great deal. Especially after wasting days dealing with the complexity of Vercel’s AI SDK, which doesn’t address the use case of tools using assistant streaming — sometimes it’s just better going straight to the source (OpenAI, in this case).
Anyhow, thank you! ✨ Martin