Get messages in real time from GPT as well as the other humans talking in the group chat, along with their online/offline and rate limit status, through federated subscriptions.
On the backend, reliably connect to OpenAI and maintain chat group state without a database through durable functions, where a function can run forever, and we can trust that the function's variables will have the correct data, even if the process running the function is restarted.
Speaker:
Loren Sands-Ramshaw
Support Engineer at Netflix