Are Assistant APIs the same as GPTs in the UI?
Trial and error development iteration is really slow on the UI, can I use Assistant API and just copy-paste the final prompt, etc. into a GPT?
What’s puzzling me is the function calling: Assistant API requires a JSON like structure whereas custom GPTs require an OpenAPI spec.
Are they the same, just with slightly different structure?
3 Likes
I just have a json file that specifies a url to call. in there.
I have a local host and forwarded with ngrok for testing via a flask api set up, and it’s calling functions pretty well.
( new developer, please don’t roast any misuse of jargon and just correct me :))
So it think if you have a json in your assistant, you could just copy/paste it into a GPT action.
onjas
4
Somehow the api.example-weather-app.com example in the GPT action don’t work, and in testing it just keeps failing.
I wonder if there’s any working example out there?
Besides GPTs and Assistant API are targeted to different people(non-technical or technical guys).
Assistant API does not have the browsing capability, and does not integrate Dall-e.
However, you can integrate web browsing and Dall-e by function calling. Here are 7 awesome assistant api demos on github, davideuler/awesome-assistant-api