Create Personal Data Vector Store from Google Sheets with AI (OpenAI/Gemini)

This n8n workflow automates the process of extracting personal data from a Google Sheet, embedding it using an AI model (OpenAI or Google Gemini), and storing it in a Supabase vector store. This enables you to create a searchable and queryable knowledge base of your personal data, which can then be used for various AI applications.

What it does

This workflow performs the following key steps:

1. Reads Data from Google Sheets: It fetches data from a specified Google Sheet, likely containing personal information or documents.
2. Converts Data to File: The retrieved Google Sheet data is converted into a file format suitable for document loading.
3. Loads Documents: A default data loader processes the converted file to extract document content.
4. Embeds Documents: It uses either OpenAI Embeddings or Google Gemini Chat Model (acting as an embedding provider in this context) to generate vector embeddings for the document content.
5. Stores in Supabase Vector Store: The generated embeddings, along with the original document content, are stored in a Supabase vector database, making them searchable.
6. Manages Chat Memory (Optional): A Postgres Chat Memory node is included, suggesting potential for conversational AI applications that leverage this vector store.
7. Triggers on Chat Message (Optional): A Chat Trigger node indicates that this workflow could be initiated by a chat message, allowing for interactive use of the personal data vector store.

Prerequisites/Requirements

To use this workflow, you will need:

• n8n Account: A running instance of n8n.
• Google Sheets Account: Access to the Google Sheet containing the personal data.
• OpenAI API Key OR Google Gemini API Key: For generating embeddings. You will need to configure credentials for the chosen AI provider.
• Supabase Account: A Supabase project with a configured vector store.
• PostgreSQL Database (Optional): If you intend to use the Postgres Chat Memory, you'll need a PostgreSQL database.

Setup/Usage

1. Import the Workflow: Download the provided JSON and import it into your n8n instance.
2. Configure Google Sheets Node:
   • Set up your Google Sheets credentials.
   • Specify the Spreadsheet ID and the range or sheet name where your personal data is located.
3. Configure Embeddings Node:
   • Choose between "Embeddings OpenAI" or "Google Gemini Chat Model" for your embeddings.
   • For OpenAI: Set up your OpenAI API Key credential.
   • For Google Gemini: Set up your Google Gemini API Key credential.
4. Configure Supabase Vector Store Node:
   • Set up your Supabase credentials (e.g., Supabase URL and API Key).
   • Specify the table name in your Supabase project where the vectors should be stored.
5. Configure Postgres Chat Memory (Optional): If you plan to use this, set up your PostgreSQL database credentials and table.
6. Activate the Workflow: Once all credentials and configurations are set, activate the workflow.

The workflow can be triggered manually to initially populate your vector store or configured to run on a schedule to keep your personal data vector store updated. The "Chat Trigger" suggests it can also be integrated into a larger conversational AI system.