Q: Hi, Is possible created a store database ?
I want to create record tables for a store, focused on products. Each product should have a name, price, and description, and it can also be related to between one and four images.
The idea is that, with this database structured in tables (and using folders to store the images), I can create workflows where an AI agent retrieves the product information along with one or more of its images, depending on what is requested.
For example, I could ask for a list of products priced between 5 and 10 dollars, and receive as a result a table with the products and their respective images.
Additionally, I would like the AI to be able to analyze an image of a product and, based on that image, automatically retrieve the corresponding record from the database.
is possible to ask to google drive or databases?
SeanP_AgenticFlowAI
Jun 20, 2025A: Hey JasonGCI!
Yes, absolutely! What you're describing is a great use case for AgenticFlow. Here's the short version:
1. Create Product Database (Tables & Images):
Tables: Yes. Use AgenticFlow's "Table" dataset feature. Create a table with columns for Product Name, Price, Description, and Image URLs (you can have columns like Image_URL_1, Image_URL_2, etc.).
Image Storage: Store your images in a cloud service like Google Drive, OneDrive, Dropbox, or even a publicly accessible S3 bucket. Put the direct shareable URLs of these images into your AgenticFlow product table.
2. AI Agent Retrieves Product Info & Images:
Yes. Create an Agent. Enable "Knowledge" on your product Table dataset.
Instruct Agent: "List products between $5 and $10." The agent will query your table and can be prompted to return the name, price, description, and the image URLs. The chat interface can then display these images if the URLs are valid.
3. AI Analyzes Image & Retrieves Record:
Yes (with a multi-step approach):
- User provides/uploads an image to the Agent.
- The Agent uses an image analysis model (like our built-in Gemini Vision capabilities, or BYOK for GPT-4V/Claude Vision) to describe the key features of the product in the image.
- The Agent then uses that text description to query your product Table dataset (using its knowledge of the 'Description' or 'Name' column) to find the matching product record.
4. Ask Google Drive or Databases:
Google Drive: Yes, via the Google Drive MCP (https://agenticflow.ai/mcp/google_drive), an agent can list files, download them, and (if text-based like a Google Doc/Sheet containing product info) extract data.
Databases (SQL, etc.): Yes, via MCPs for specific databases like Supabase (https://agenticflow.ai/mcp/supabase), MySQL, PostgreSQL, etc., or by using the API Call node if your database has a REST API. The agent can then query these directly.
So, your vision for a product database with AI-driven querying (text-based and image-based) is definitely achievable with AgenticFlow by combining Table datasets, image storage, LLMs for querying/analysis, and vision models!
— Sean