Hi glad to hear it helped. It would be similar in that you're setting up API calls like shown but there's additional data management to handle. Meaning, each new message in the conversation is added to the bundle of messages sent over to the API every time you type something new. That way, it can keep track of the evolving conversation. A data configuration that is setup to handle this is available in the full chat configuration resource in the video description. But for full disclosure (1) that resource's intention is more about being able to copy+paste to get up and running in a chat-based UI experience rather than attempt to explain each detail of the configuration and (2) that resource is not for the assistant API but that shouldn't matter because the data structure is the same, the only difference with assistant API is you're getting the extra fine tuning that comes along with building an assistant vs the out-of-the-box general LLM. Hope that helps.
Hey indeed! If you will, not them or a nebulous "everyone" though, you and me! The empowerment to build something awesome begins with each of us. Pretty exciting stuff!
dude you are amazing!! one of the few recent vids I can find on this. thank you thank you thank you
Glad I could help!
this is really helpful and well explained. what if instead of completions API you wanted to use an assistant API?
Hi glad to hear it helped. It would be similar in that you're setting up API calls like shown but there's additional data management to handle. Meaning, each new message in the conversation is added to the bundle of messages sent over to the API every time you type something new. That way, it can keep track of the evolving conversation.
A data configuration that is setup to handle this is available in the full chat configuration resource in the video description. But for full disclosure (1) that resource's intention is more about being able to copy+paste to get up and running in a chat-based UI experience rather than attempt to explain each detail of the configuration and (2) that resource is not for the assistant API but that shouldn't matter because the data structure is the same, the only difference with assistant API is you're getting the extra fine tuning that comes along with building an assistant vs the out-of-the-box general LLM. Hope that helps.
Awesome!! Thank you for explaining this 👊
You're very welcome
Amazon features!!! Wow everyone will be able to build anything that comes from their mind.
Hey indeed! If you will, not them or a nebulous "everyone" though, you and me! The empowerment to build something awesome begins with each of us. Pretty exciting stuff!
Subscribed! You look like you could be my brother or cousin!
Appreciate that! Must be a couple of handsome guys 😆.