Curious what you use AI for in relation to bubble?
I’m bad at UI/UX so I’ve gotten chat gpt help by sharing screenshots and it gives me some basic tips. Definitely not nearly as good as hiring someone, but 100x better than me not using it
Curious what you use AI for in relation to bubble?
I’m bad at UI/UX so I’ve gotten chat gpt help by sharing screenshots and it gives me some basic tips. Definitely not nearly as good as hiring someone, but 100x better than me not using it
At first I thought you were asking how are we integrating AI into the applications we are building, but reading your comment, it sounds like you are asking how are we using AI to help us build a Bubble app.
For what I think you are really asking (ie: how am I using AI to help me build a Bubble app)…not at all.
For what I thought the question was about (ie: how am I integrating AI into the apps I build)…
I’d be interested to hear what other people are using AI to integrate into the apps they are building.
How do you do that? Do you pass the AI a list of options then use the response to run a conditionally run a specific event?
Sort of.
Give the AI a prompt that indicates when the AI responds to structure that response in such a way that it has values that are utilized by conditionals to run specific events.
So if user says something like, ‘yes, please buy that item for me’ the AI would respond with something like ‘okay, I’ll place the order now. {log:command:function:blahblahblah}’.
Have not implemented it for an app wide functionality, but put it together to automatically send data via API to another database based on user responses to questions from the AI.
Interesting. Would you need to convert the voice prompt to text first then pass that to the AI? The response of the AI (text) would then be converted to audio?
Yes
Yes
Both! I haven’t started integrations yet and have been wondering if it’s worth it.
Is this possible ? .Your user writes ‘’ I would like to have a cat ‘’ in the searchbox , it goes to AI and converts that reccomended input to actionable result in your app.
I don’t even know what that should do, how would AI?!
It allows for most convenient user experience , is this possible with AI I am wondering , or are there easier ways to do this
Well you can get AI to create a new Thing, sure… different specifics make it more complex but it’s perfectly possible.
no no , I mean there is a cats section on my app already and a separate filter system for cats already existent.I want my user to write whatever he wants to , such as ‘‘I want to buy a cat , that is grey colored’’ , and searchbox outputs the cats page with scottish kind cat filtered.Assume I have already created the database and everything. This type of functionality
Disclaimer : This is an example
I think if you’re already organizing the information like this you wouldn’t need an AI?
A few underrated ways I use GPT4:
Creating option sets. It’s hard to get an option set to be all encompassing and concise. Ex: If I want users to select what type of home they live in, I use it to create an option set for all types of home styles.
Image ranking and moderation. I use GPT4 with vision to rank images that are uploaded to my app based on some criteria. Then I use that ranking to dynamically sort images on profiles. So the best images are generally always first, without human intervention.
Brainstorming database schemas. If I’m trying to add a new feature like allowing users to friend each other. GPT4 is able to recommend to me how I can structure the necessary tables and recommend what fields should be in the datatype.
Pasting errors from console or Bubble logs to help me debug or recommend different approaches.
A complete replacement for this forum. Note, this only works because I have a deep understanding of Bubble and its features. So I’m able to sniff out if Bubble advice it’s giving me is hallucinating or not. But I have to say that GPT4 is MUCH better at Bubble questions than the vanilla GPT3.5.
Going from 3.5 to 4 meant going from playing with AI to actually using it on the daily.
Can you elaborate on that ?
A user uploads an image → I create a “Media Container” which stores the image and some number fields. One of those fields is “Overall rating”.
I have a database trigger that fires when a Media container is created.
That trigger sends an API call to OpenAI with the uploaded image. My API call has a system prompt that lays out the criteria for rating the image and the JSON format I’m expecting back.
I then parse the JSON response and add all the image ratings to the Media Container.
I also tell GPT4 to give the image an Overall rating of 0 if it is NSFW.
In my app’s repeating group of “Media containers”, I sort them based on the Overall rating and only show Media containers that have a rating > 0.
That is great tech , how much it costs and do you have API docs link on this ?
Pricing isn’t cheap for processing images. But for my use case, it’s worth it and all OpenAI products seem to get cheaper over time. So I’m not stressing about the cost right now.