In today's world, no-code platforms like Bubble have made it possible to build complex applications without writing a single line of code. One exciting area where these platforms can be leveraged is in the realm of artificial intelligence (AI). In this post, we'll take you step-by-step through the process of creating an AI-driven application on Bubble, utilizing AI embeddings.

What are Embeddings?

In simple terms, embeddings are mathematical representations of text in a high-dimensional space. AI uses these embeddings to understand the semantic meaning of the text, and it is this understanding that allows AI to find patterns, similarities, and differences in text data.

Why use Embeddings?

Embeddings are particularly useful when dealing with large amounts of text. For instance, if you have a whole PDF, you would need to split that up into chunks in order to process those chunks into a vector database, such as Pinecone. This is where embeddings come in – they allow you to chunk the text and then store the embedded text in the vector database.

Getting Started with Bubble

To get started, create an account on Bubble and navigate to your dashboard. Click on 'Create an app' at the top right and name your application something unique. For this tutorial, we're using 'embeddings demo no code'.

Note that your Bubble app will need to be on a paid account or a free trial account, as there's a specific requirement in this tutorial to chunk text into smaller groups, a feature not available in the free version. Bubble usually offers a free trial for two weeks which should suffice for this tutorial.

Setting up the API connections

Once you have set up your app, the first thing you need to do is to establish some API connections, which will allow you to connect to external services such as OpenAI and Pinecone.

First, add the API connector plugin to your Bubble app. Then, set up your OpenAI connection by adding a new API and naming it OpenAI. You will need to provide your OpenAI API keys and other details to establish the connection.

The next step is to set up an API call for embeddings using OpenAI. Here, you will send a piece of text to OpenAI and get a response containing the embedded text.

Finally, you will need to set up a Pinecone API call. Pinecone is the vector database where you will be storing those chunks of text.

This guide should serve as a handy reference as you venture into the exciting world of no-code AI application development on Bubble.

To learn how to build the entire AI chatbot app with AI embeddings, check out our full course here.