How to build a book recommendation site with AI

In today’s digital age, leveraging AI and NoCode tools can streamline many tasks, including building custom landing pages. Recently, I dove into the world of AI to create a website that generates personalized landing pages for my favorite books. Here’s how I did it.

AI and NoCode: The Tools

To bring this project to life, I used several key tools:

• [GPT-3.5 by OpenAI](): For generating content.
• Webflow: For designing and hosting the website.
• Airtable: To manage the data and structure the database.
• WhaleSync: To sync data between Airtable and Webflow.

These tools, combined with some creativity, made it possible to create a seamless workflow for generating and updating landing pages.

Step-by-Step Process

1. Concept and Structure

The concept behind booknotes.ai was simple: create a landing page for each book that provides an overview, key takeaways, and a link to purchase the book on Amazon. The first step was to structure the database in Airtable. Airtable's flexibility allowed me to create a comprehensive database with fields for the book’s title, author, genre, summary, key takeaways, and target audience.

[Description of the image of the Airtable database structure]

2. Content Generation

Using GPT-3.5, I generated summaries and key takeaways for each book. The workflow was set up in Airtable and automated using Make.com. Here’s a breakdown:

• Database Setup: Created fields in Airtable for all necessary information.
• Book Selection: Picked books and added them to Airtable using a web clipper for Amazon.
• Summary Generation: Triggered an API request to GPT-3.5 to generate summaries and key takeaways.
• Review and Edit: Manually reviewed the generated content to ensure accuracy and relevance.

‍

3. Integration with Webflow

Next, I needed to display this data on a website. WhaleSync was used to synchronize data between Airtable and Webflow. The process involved:

• Setting Up Collections: Created collections in Webflow corresponding to the Airtable fields.
• Mapping Fields: Used WhaleSync to map Airtable fields to Webflow collections.
• Website Layout: Designed a simple yet effective website using Re-loom components, ensuring a clean and engaging user interface.

‍

4. Automation and Efficiency

Automation was key to making this project efficient. From content generation to website updates, everything was streamlined:

• Automated Updates: Once a book's information was ready, it was automatically sent to Webflow.
• Real-Time Sync: WhaleSync ensured that any updates in Airtable were reflected on the website instantly.

‍

AI API and Costs

Using the OpenAI API for content generation is cost-effective. The pricing is based on a token model, with each 1,000 tokens costing $0.02. For example, generating a summary for a book like "The 4-Hour Workweek" cost only $0.02. This affordability makes it feasible to scale the project without significant expenses.

Challenges and Feedback

While the project was successful, it wasn’t without challenges:

• Response Quality: Some responses from the API required fine-tuning of prompts.
• Formatting Issues: Minor formatting issues were resolved using NoCode tools.
• Experimentation: Continuous experimentation with prompts was necessary to achieve optimal results.

Overall, the experience was enlightening. The combination of AI and NoCode tools opens up numerous possibilities for creating automated, dynamic websites with minimal effort.

Conclusion

This project highlighted the power of AI and NoCode tools in creating efficient workflows for content generation and website management. If you’re looking to explore similar projects, don’t be intimidated by the technology—these tools are more accessible than ever.

For more detailed tutorials and to stay updated, subscribe to my newsletter. If you enjoyed this post, like, subscribe, and let me know in the comments what AI projects you're working on.

‍