Watch Me Build a $7.5k AI Price Tracker in 40 min (N8N) (No-Code)

By Ben AI

TechnologyAIBusiness
Share:

Key Concepts

  • No-code AI competitor price tracking: Building a system without coding to monitor competitor prices using AI.
  • Web scraping: Extracting data from websites.
  • Anti-scraping protection: Measures implemented by websites to prevent automated data extraction.
  • Naden: A no-code automation platform with AI capabilities.
  • Lucer Studio: A data visualization tool for creating dashboards.
  • LLM (Language Model): An AI model used for tasks like price extraction and comparison.
  • Structured output: Formatting LLM output into a specific structure (e.g., JSON) for easy use.
  • API (Application Programming Interface): A set of rules and specifications that software programs can follow to communicate with each other.

System Overview and Demo

The video demonstrates a no-code AI system built in Naden to track competitor prices for an e-commerce brand. The system:

  • Scrapes prices from competitor websites multiple times a day.
  • Displays the prices in a Lucer Studio dashboard.
  • Notifies the company via Slack when a competitor has a lower price.

The demo uses Nike as the example brand and tracks prices against two competitors: Sporting Goods and Finish Line. The system uses Google Sheets to store product URLs, names, and prices for Nike and its competitors. The Naden workflow scrapes these websites, extracts prices using LLMs, updates the Google Sheets, and compares prices, sending Slack notifications when a competitor's price is lower.

Building the System in Naden: Step-by-Step

1. Trigger Setup

  • A "Schedule" trigger is used to run the workflow twice a day (every 12 hours).

2. Scraping Nike (Our Brand)

  • Get Rows (Google Sheets): Retrieves product URLs, names, and prices from the Nike Google Sheet.
  • HTTP Request: Scrapes the HTML content from the Nike product URLs. This works because Nike's website doesn't have strong anti-scraping measures.
  • Markdown: Cleans up the HTML data by removing unnecessary elements like image tags (<img>).
  • LLM (Language Model): Extracts the price from the cleaned HTML using Google Gemini (2.0 Flash, free experimental version). A system prompt instructs the LLM to extract only the price. Structured output is used to format the output as a JSON object with a "price" field.
  • Update a Row (Google Sheets): Updates the Nike Google Sheet with the extracted prices.

3. Scraping Competitor Websites (Sporting Goods, Finish Line)

  • Get Rows (Google Sheets): Retrieves product URLs, names, and prices from the competitor's Google Sheet (e.g., Sporting Goods).
  • Split: Splits the list of product URLs into individual URLs. This is done to avoid being blocked by anti-scraping measures.
  • Loop Over Items: Creates a loop to process each product URL separately.
  • Web Scraper (ScrappyFly API): Uses the ScrappyFly API to scrape the HTML content from the competitor's product URLs. This is necessary because competitor websites have anti-scraping protection.
    • Curl: A curl request is used to interact with the ScrappyFly API. The URL and query parameters (including the target URL and anti-scraping settings) are configured.
    • Map: Extracts only the HTML content from the ScrappyFly API response.
  • Markdown: Cleans up the HTML data.
  • LLM (Language Model): Extracts the price from the cleaned HTML using Google Gemini.
  • Update a Row (Google Sheets): Updates the competitor's Google Sheet with the extracted price.

4. Price Comparison and Slack Notification

  • Get Rows (Google Sheets): Retrieves the corresponding Nike product price from the Nike Google Sheet, filtering by product name.
  • LLM (Language Model): Compares the competitor's price with the Nike price. The system prompt instructs the LLM to output "lower" only if the competitor's price is lower than Nike's price.
  • If: A conditional statement checks the output of the price comparison LLM.
  • Send a Slack Message: If the competitor's price is lower, a Slack notification is sent to a specified channel.
  • Loop Back: The workflow loops back to process the next product URL.

5. Repeating for Other Competitors

The entire competitor scraping, price comparison, and notification process is duplicated for each additional competitor.

Web Scraping Techniques

The video highlights two web scraping methods:

  • Simple HTTP Request: Suitable for websites without anti-scraping protection.
  • Advanced Web Scraping API (ScrappyFly): Necessary for websites with anti-scraping protection. ScrappyFly offers features like:
    • Automatic anti-bot bypass
    • Proxy rotation
    • JavaScript rendering
    • Geolocation selection
    • Delay before extracting content

Building a Lucer Studio Dashboard

The video demonstrates how to create a Lucer Studio dashboard to visualize the price data:

  1. Connect Data Sources: Connect Lucer Studio to the Google Sheets containing the price data for Nike and its competitors.
  2. Create Blended Data: Combine the data from the separate Google Sheets into a single blended data source. This allows you to represent all the data in a single chart or table.
    • Join Conditions: Configure the join conditions to map the data correctly. In this case, the product names are used to join the tables.
    • Inner Join: Use an inner join to only include products that are available in all tables.
  3. Add Charts and Tables: Add charts (e.g., bar charts) and tables to the dashboard to visualize the price data.
    • Dimensions and Metrics: Configure the dimensions (e.g., product name) and metrics (e.g., price) for each chart or table.
    • Calculated Fields: Add calculated fields to perform calculations on the data (e.g., calculate the price difference between Nike and its competitors).
  4. Customize the Dashboard: Customize the style and layout of the dashboard to make it visually appealing and easy to understand.

Naden Advantages

The video emphasizes the following advantages of using Naden:

  • Combination of Workflow Automation and AI Agents: Naden combines strong logic-based automation capabilities with AI agent functionality.
  • Cost-Efficiency: Naden is generally cheaper to run than other no-code automation platforms.
  • Customization Options: Naden offers more customization options compared to some other platforms.

Conclusion

The video provides a detailed, step-by-step guide to building a no-code AI competitor price tracking system using Naden and Lucer Studio. It covers web scraping techniques, LLM integration, data visualization, and the advantages of using Naden. The system automates the process of monitoring competitor prices, enabling e-commerce businesses to maintain competitive pricing strategies. The use of ScrappyFly API is crucial for scraping websites with anti-scraping measures, and Lucer Studio provides a user-friendly way to visualize the data.

Chat with this Video

AI-Powered

Hi! I can answer questions about this video "Watch Me Build a $7.5k AI Price Tracker in 40 min (N8N) (No-Code)". What would you like to know?

Chat is based on the transcript of this video and may not be 100% accurate.

Related Videos

Ready to summarize another video?

Summarize YouTube Video