How integrated is CSAI with Cornerstone itself?

Support
#1

I’ve seen the CSAI YouTube update, and while it looks pretty useful to have ChatGPT etc within Cornerstone rather than going off to a browser, is the AI element of Cornerstone actually built with Cornerstone in mind - or is it basically embedded into the builder? For example, if I ask it to do something specific to my build, like "create 3 rows with 4 columns each, can this be achieved, or are its uses limited to creating code to add into the page CSS for example, as ChatGPT and others would already do? From what I saw on the YouTube video, it’s basically doing the same as it would online, apart from reducing the copy/paste processes by a few clicks…

I’m just trying to ascertain its uses before purchasing or not.

Thanks!

#2

It’s very integrated and will only get more so. Generating layouts in the manner you described will be coming with our own ThemecoAI provider which is actively being trained. The base of our vision is bringing the power of the many different AI models/providers directly to the Cornerstone experience to help with specific actions like customizing Elements, fine tuning images, generating audio/video, or even hosting your own AI to name a few.

#3

Awesome news! Will your ThemecoAI be able to generate the correct shemantics/tags/w3c/wacg standard elements/layouts?

#4

Some of that has to do with our Elements which already try to have proper semantics. Our training will include how to change html tags, so it should already try to do this when we release it. As with all things AI, it’s heavily influenced on the training data. Cornerstone has a very deep bench to pull from in this regard.

#5

So just checking, ThemecoAI will be something like ChatGPT/Claude but specifically for Cornerstone OR we’re still expected to add our own API key to use ThemecoAI?

And how would the current CSAI one time payment fit into this plan?

#6

Correct, it will be a separate provider option just like ChatGPT, Claude, etc. There are hard costs on a per request basis to run and host these models so not sure how the pricing will look just yet.