Extract Amazon Product Data with Scraped HTML using GPT-4 and Google Sheets

This n8n workflow automates the process of extracting structured product data from Amazon product pages (provided as HTML), enriching it with GPT-4, and then storing the results in Google Sheets. It's designed to streamline data collection for market research, competitor analysis, or inventory management.

What it does

1. Triggers Manually: The workflow starts when you manually execute it within n8n.
2. Reads HTML Input: It expects HTML content (likely from an Amazon product page) as input.
3. Extracts Product Details (HTML): Uses an HTML node to parse the provided HTML and extract specific product information (e.g., product title, price, description).
4. Splits Data for Processing: If multiple items are processed, it splits them into individual batches for iterative handling.
5. Enriches Data with GPT-4: For each product, it uses a LangChain Basic LLM Chain with an OpenAI Chat Model and a Structured Output Parser to intelligently extract and structure specific product attributes (e.g., features, specifications, benefits) from the raw HTML or previously extracted data. This leverages the advanced understanding capabilities of GPT-4.
6. Saves to Google Sheets: Finally, it appends the extracted and enriched product data to a specified Google Sheet.

Prerequisites/Requirements

• n8n Instance: A running n8n instance (cloud or self-hosted).
• OpenAI API Key: An API key for OpenAI, with access to GPT-4 models, configured as an n8n credential.
• Google Sheets Account: A Google account with access to Google Sheets, configured as an n8n credential.
• Amazon Product HTML: The workflow expects raw HTML content of Amazon product pages as input. This typically comes from a previous step (e.g., an HTTP Request node for scraping, or manual input for testing).
• Scraped Amazon Product Data: The workflow assumes you have a method to obtain the raw HTML of Amazon product pages. While this workflow processes the HTML, it doesn't scrape it directly. An HTTP Request node is present, but not connected as a primary data source in the provided JSON, suggesting it might be used for testing or an alternative input path.

Setup/Usage

1. Import the Workflow:
   • Copy the provided JSON code.
   • In your n8n instance, go to "Workflows" and click "New".
   • Click the "Import from JSON" button and paste the workflow JSON.
2. Configure Credentials:
   • OpenAI Chat Model: Click on the "OpenAI Chat Model" node. Select or create an "OpenAI API" credential. Ensure it has access to the models you intend to use (e.g., gpt-4).
   • Google Sheets: Click on the "Google Sheets" node. Select or create a "Google Sheets" credential. Grant it the necessary permissions to write to your desired spreadsheet.
3. Configure Nodes:
   • HTML (Extract Product Details): Review the CSS selectors in this node to ensure they correctly target the product title, description, and other elements you wish to extract from the Amazon HTML. You may need to adjust these based on Amazon's current HTML structure.
   • Basic LLM Chain:
     • Prompt: Adjust the prompt within this node to guide GPT-4 on what specific information to extract and how to structure it.
     • Output Parser: Review the schema in the "Structured Output Parser" node. This defines the JSON structure that GPT-4 should return. Modify it to match the exact fields you want to extract (e.g., product_name, price, features_list, category).
   • Google Sheets:
     • Spreadsheet ID: Enter the ID of your Google Sheet.
     • Sheet Name: Specify the name of the sheet within the spreadsheet where data should be appended.
     • Column Mapping: Ensure the column mapping aligns with the output from the Basic LLM Chain and the columns in your Google Sheet.
4. Provide Input HTML:
   • For testing, you can manually input sample Amazon product HTML into the first node (e.g., using a "Set" node before the "HTML" node, or by directly pasting it into the "HTTP Request" node if you configure it to accept static data).
   • For live use, you would typically connect a preceding node (e.g., an HTTP Request node configured to scrape an Amazon URL, or a Webhook receiving HTML) to feed the Amazon product page HTML into the "HTML" node.
5. Execute the Workflow: Click "Execute Workflow" to run it. Observe the output at each step to ensure data is being processed as expected.

This workflow provides a powerful foundation for automating product data extraction and enrichment, leveraging the capabilities of large language models for intelligent data structuring.