Overview

From chatbots to large language models it seems AI is everywhere. Developer tools and platforms are moving at an incredible pace making it possible for anyone with some coding skills to build highly complex AI-driven apps in a short amount of time. At the heart of any AI application is access to the right data. The Airbyte platform is a core aspect of building the right AI data stack to unlock data, whether it is structured or unstructed and make it available for AI use cases.

In this tutorial, you will build an AI-powered chatbot to allow users to interact with e-commerce data. They will be able to ask natural language questions to uncover insights in the data.

What You Will Learn

In this tutorial, you will learn the following how to deploy, configure, and create an AI+data full stack application.

You will get hands on with:

• Airbyte Cloud to connect to Stripe test data
• Use the Airbyte Postgres Destination connector to send Stripe data to Postgres, deployed on Supabase
• Configure Supabase to use PGVector to support embeddings
• Create a data pipeline in Airbyte to handle sync tasks and send data embeddings for AI use cases
• Create SQL functions to work with openAI question embeddings
• Write python-based chatbot uses OpenAI APIs to interact with your data and embeddings.
• (bonus) Create a full-stack web application with a frontend in Next.js, to host your chatbot

Whilst a basic understanding of coding in Next.js, Python, and SQL is helpful, if you are not comfortable with coding in these languages, don’t worry! All of the code will be provided for you throughout.