This AI System Finds Amazon's Top Products on AUTOPILOT (No-Code n8n Tutorial)

Key Concepts:

• No-Code Automation: Building automated workflows without writing code.
• Web Scraping: Extracting data from websites.
• Bright Data: A web data platform used for proxy and web scraping services.
• NADN: A no-code automation platform.
• OpenAI Node: Using OpenAI's models (like GPT-4) for text summarization and data extraction.
• Google Sheets Integration: Outputting scraped data to a Google Sheet.
• Data Extraction Schema: Defining the structure of data to be extracted by AI.
• HTTP Request Node: Sending requests to external APIs (like Bright Data).
• Split Item Node: Splitting a list of items into individual data rows.

1. Introduction

• The video demonstrates how to automate the process of identifying top-selling products on Amazon using AI and no-code tools.
• The goal is to create a system that automatically sends daily updates on top-selling products in specific niches (e.g., home improvement, electronics) to the user's email or Google Sheet.
• The solution is built using NADN, a no-code platform, making it accessible to users without programming experience.

2. Overview of the Automated Workflow

• The workflow involves scraping Amazon product pages, extracting relevant data, summarizing it using AI, and outputting the results to a Google Sheet and/or a Slack channel.
• The process is designed to be fully automated, running daily without manual intervention once set up.

3. Setting Up the NADN Workflow

• Creating a NADN Account: Users are directed to sign up for a free 14-day trial of NADN via a link in the video description.
• Importing a Blueprint (JSON File): The video demonstrates how to import a pre-built workflow blueprint (JSON file) into NADN to quickly set up the automation. This blueprint contains all the necessary nodes and configurations.
  • The blueprint is available to members of the creator's community.
• Manual vs. Scheduled Trigger: The video uses a manual trigger ("When clicking a test workflow") for demonstration purposes. For production use, a scheduled trigger is recommended to automate the process daily.

4. Web Scraping with Bright Data

• Bright Data Account Setup: Users are instructed to create a Bright Data account to access web scraping tools. A free trial is available.
• Creating a Web Unlocker Zone: Within Bright Data, a "Web Unlocker API" zone is created. This zone allows the user to scrape websites without being blocked.
  • The zone requires a name (lowercase) and generates an API token.
• Set Node Configuration: A "Set" node in NADN is used to define variables:
  • zone: Stores the name of the Bright Data Web Unlocker zone.
  • URL: Stores the Amazon URL to be scraped (e.g., the URL for "Bestsellers in Home & Kitchen").
• HTTP Request Node Configuration: An "HTTP Request" node is configured to send a POST request to the Bright Data API:
  • Method: POST
  • URL: api.brightdata.com/request
  • Authentication: Header with "Authorization" set to "Bearer [API Token]"
  • Body:
    • zone: The Bright Data zone name.
    • URL: The Amazon URL to scrape.
    • format: Set to "raw".

5. Data Extraction and Summarization with OpenAI

• Code Node for HTML Sampling: A "Code" node is used to extract a sample of HTML from the scraped data. This is done to reduce the amount of data sent to the OpenAI node, improving efficiency.
  • The code (provided in the video description or community) uses JavaScript to grab a partial HTML structure.
• Information Extractor Node (OpenAI): An "Information Extractor" node is used to extract structured data from the HTML sample using OpenAI's models.
  • Model: GPT-4 (or similar) is used for its ability to understand and extract data from unstructured text.
  • Schema Definition: A schema is defined to specify the data fields to extract (e.g., title, stars, total ratings, offer, product URL, image URL). The schema is provided in the video description or community.
• Split Item Node: A "Split Item" node is used to split the output from the OpenAI node into individual items (rows), making it easier to process and output to a Google Sheet.

6. Outputting Data to Google Sheets

• Google Sheets Node Configuration: A "Google Sheets" node is configured to append or update rows in a Google Sheet.
  • Credentials: Google Sheets credentials are set up to allow NADN to access the sheet.
  • Resource: "Sheet within document"
  • Operation: "Append or update row from list"
  • Document ID: The ID of the Google Sheet.
  • Sheet Name: The name of the sheet within the document (e.g., "Sheet1").
  • Column to Match On: A column to use for matching rows (e.g., "ranking").
  • Column Mapping: Data fields from the "Split Item" node are mapped to the corresponding columns in the Google Sheet (e.g., title to title, stars to stars).
• Google Sheet Template: A Google Sheet template is provided to ensure the correct column names and structure for the data.

7. Optional: Slack Integration for Summarized Updates

• Message a Model Node (OpenAI): A "Message a Model" node is used to generate a summarized update of the top-selling products using OpenAI.
  • Model: A smaller model like "nano" can be used for faster processing.
  • Prompt: A prompt is defined to instruct the AI to summarize the top 10 products.
• Slack Node: A "Slack" node is used to send the summarized update to a Slack channel or user.

8. Example and Customization

• The video demonstrates how to change the Amazon URL to scrape different product categories (e.g., clothing).
• The process is the same: update the URL in the "Set" node, and the workflow will automatically scrape and extract data for the new category.

9. Conclusion

• The video provides a step-by-step guide to automating the process of identifying top-selling products on Amazon using no-code tools and AI.
• The solution can be customized to track different product categories and output data to Google Sheets and/or Slack.
• The automated workflow saves time and effort by eliminating the need to manually monitor Amazon product pages.