OpenAI’s GPT is incredibly versatile, capable of generating human-like text across a wide range of topics and styles. Its advanced language understanding enables it to assist with tasks such as drafting emails, writing code, and even composing creative stories.
GPT’s ability to understand context and provide coherent, relevant responses makes it a powerful tool for both personal and professional use.
Additionally, its continuous improvement through iterative training ensures it stays at the cutting edge of AI technology.
As additional note, if you want to connect the AI with your Bubble database, this plugin has this built-in feature through tools/functions.
Let’s take a basic example and that you have both a Users and a Testimonials tables.
You want the AI to answer based on any requests requiring those data.
We also want to send the right data and not the full database (you may but that’s not my recommendation from token usage and context window perspective), so we must detect what is the user intent and fed the right data to the AI.
It is very simple.
You first have to define this function/tool. The below function example describes what the tool does, when it must be called and what arguments it must return (if any). Here we ask the tool, when user requests specific data, to pass to the plugin the arguments “data_type” which is either “users” or “testimonials” value. We will use this as conditional to execute the desired “Search for” to return the data to the AI.
When the AI will infer from the user query that he requires to call this tool, Call Function event start of the plugin will trigger. The element will expose the Function Name and the relevant Function Arguments value defined previously:
So you just have to use Send Function Result action and return in Function Result field either user’s or testimonials’s data depending on Function Arguments value.
by wise:able
