When im adding actions to GPT’s i barely comprehend how specs work (been at this a few months)
but… I have no idea how authentication works, or what-when-where-how.
Just tried putting in my anime list API as an experiment.
Getting errors “wrong API key”
Come on man, I just want to make a GPT that can suggest my next anime, and give it better tools than general “top 20 animes in 2023”
Like 9% of you get it, the rest are laughing.
An Action might be overkill just for getting an anime recommendation, are you sure you want your GPT to make an outbound call over the web? THe chat interface could probably do okay recommendation if you uploaded some data as a File to augment the conversation
The point is not the action - its the experiment in learning…
And lets be honest… why not?
the application might be basic but anyone who actually likes anime would love to have an AI to run through some new sh-stuff.
reguardless - Im not trying to make money off it - perhaps this thread appeals to some newbie thats trying to bite off more than he can chew, and you know what… in the era of LLM AI’s… thats viable.
If the error message is “Wrong API key” then you can double-check if the API key you are sending is the one you actually need to send.
From your description it is hard to provide support as I would not even know if the error message is comming from the OpenAI server or the server hosting your Anime library.
Maybe you can share some context and I am sure this can be dealt with.
copy pastye the API docs into GPT-4 and ask it the right question
Paste the response here.
share API endpoint ( not full url )
Like this
GET /titles/series/{seriesId}
GET /titles/search/keyword/{keyword}
we will help you to write api schema and API key