Why such a shabby experiment with the chatbot?

Dear ai2 developer team,
Experimental says, under "Model"

See chatbot/imagebot for the current list of supported models. Leaving this blank will result in the default model set by the provider being used

If you put chatbot/imagebot into your browser you get redirected to a google doc which says:

Currently, Model is for future use and does nothing.

So, I can't set the model, and you also don't tell us which model you use, so I have no idea how to get an API key. The doc is dated June 18, 2023: um, has the future perhaps arrived?

The doc also states:

If you supply your own ApiKey, which you obtain from OpenAI (we’ll add instructions to this document soon). If you have your own ApiKey, then there are no restrictions imposed by us (MIT App Inventor).

15 months ago you said "soon" How much longer will that be?

And when you update the document, maybe you could check if chatgpt does a better job of proof reading than did the author of the above cited statements.

A brilliant physics professor in New Delhi with ALS can only move his head, and his speech is unintelligible. I was hoping your chatbot could be part of a solution to let him type word by word using head movement, so please pardon the impatience.

Without an API key, your chatbot lets me send 3 requests a day. During debugging of this app, I could easily need to send 3 requests a minute.

I am a huge MIT App Inventor fan, and have used it a lot in teaching, and in free apps for persons with multiple disabilities. I would be miserable if a small matter such as this would force me to move to a platform that does a better job of allowing me to integrate AI into my apps.

warmly,
Arun Mehta

By default, both components use OpenAI.

You have a worthwhile Project.

Sign up with OpenAI to avoid the chatbot limitations.

The link below explains why the default ChatBot tool setting does not provide unlimited use of OpenAI. If you want to go beyond that usage, then you need to sign up for an OpenAI key so they can bill you accordingly.

Steve, could you please tell me which Open AI model I should get an API key for, in order to use the chatbot without restriction?
Arun

I am not sure it is possible to select the AI model using the ChatBot tool yet. I think it can only be used with the model the tool defaults. I have only used the basic model default in the tool. It works fine. I realize you want a more robust model. To use more sophisticated models I think MIT needs to update the component.

A possible extension alternative to using the native ChatBot tool you might be able to use for your Project is

[F/OS] 🧠 Artificial Intelligence and OpenAI! .I think the extension only works with gpt.3 models and lower. Read the entire link or ask your question in that thread. Sorry I can't be of more help Arun.

and some information about pricing

https://openai.com/api/pricing/

Here are some community discussions about using chat bot technology with App inventor that might be useful.