How to Automate Thesis Generation Using Dumpling AI and Make.com

Generating a well-structured thesis requires extensive research, filtering through large amounts of information, and organizing key insights into a logical format. This can be a time-consuming process—especially when dealing with multiple sources.

With Make.com and Dumpling AI, you can fully automate thesis generation by scraping content from multiple websites, filtering relevant information, and generating a well-structured thesis using AI. The final output is then stored in Airtable for easy access.

This tutorial will guide you step by step through setting up an automation that dynamically collects, processes, and organizes research data into a structured thesis. Whether the topic is climate change, artificial intelligence, finance, or medicine, this workflow adapts to any subject.

Benefits of This Automation

✅ Saves Research Time – Automatically gathers and filters content from multiple sources.
✅ Ensures Quality – AI extracts only the most relevant data.
✅ Provides Structure – Thesis follows a predefined, professional format.
✅ Scalable & Adaptable – Works for any topic, not just climate change.
✅ Easy to Retrieve & Manage – Final content is stored in Airtable for future use.

Step 1: Crawl Websites for Relevant Content

This module extracts structured content from the target website(s) and prepares it for further processing.

The first step is to crawl web pages to collect relevant information. We use Dumpling AI’s crawler module in Make.com for this.

Configuration:

• Module: Dumpling AI – Crawl
• URL: [Enter the website URL] (Example: https://www.un.org/en/climatechange/what-is-climate-change).
• Depth: 2 (Defines how deep the crawler should go).
• Limit: 10 (Max number of pages to crawl).
• Format: Markdown (Structured content extraction).

Step 2: Feed Extracted Data into an Iterator

Since we are crawling multiple pages, we need to process each extracted page separately.

Configuration:

• Module: Iterator
• Input: The array of extracted data from the previous step.

What This Does:

The iterator processes each extracted page separately, allowing us to scrape and refine the content dynamically.

Step 3: Scrape and Clean the Extracted Data

Now that we have extracted URLs, we scrape the actual content from those links while ensuring it’s cleaned and formatted properly.

Configuration:

• Module: Dumpling AI – Scrape
• URL: {{2.url}} (Dynamically sourced from Step 2).
• Format: Markdown.
• Cleaned: True (Removes unnecessary elements like ads and unrelated sections).
• Render JavaScript: True (Ensures dynamically loaded content is also captured).

What This Does:

This step extracts the core text from each URL, ensuring clean, structured content.

Step 4: Aggregate Scraped Content

Since we may have multiple pages of extracted content, we need to combine them into a single structured text before sending it to AI.

Configuration:

• Module: Text Aggregator
• Row Separator: Empty (Ensures smooth text flow).
• Value: {{3.content}} (Merges all scraped content).

What This Does:

This step compiles all extracted data into one unified text block for better AI processing.

Step 5: Use AI to Generate a Structured Thesis

Now, we use OpenAI’s GPT-4 model to process the extracted data and generate a well-structured thesis.

Configuration:

• Module: OpenAI – Create Completion
• Model: ChatGPT-4o-latest.
• Temperature: 1 (Higher creativity).
• Max Tokens: 2048 (Ensures a detailed response).
• System Instructions:
  • Filter out irrelevant content.
  • Organize extracted data into a structured thesis using this format:
    • Introduction
    • Key Issues & Causes
    • Impacts
    • Mitigation & Solutions
    • Conclusion
• User Input: The aggregated content from Step 4.

What This Does:

AI analyzes and filters the extracted content to produce a structured thesis based on the topic.

Step 6: Store the Thesis in Airtable

Once the AI generates the thesis, we store it in Airtable for easy tracking and retrieval.

Configuration:

• Module: Airtable – Create Record
• Base: My Thesis.
• Table: Thesis.
• Record Field: Thesis → {{5.result}} (The AI-generated thesis).

What This Does:

This step saves the thesis in Airtable, making it easy to manage, retrieve, and update when needed.

Conclusion

This fully automated workflow allows you to generate a well-structured thesis on any topic by:

1. Scraping relevant information from websites.
2. Filtering the data to remove unnecessary content.
3. Using AI to structure the thesis in a professional format.
4. Storing the final output in Airtable for easy access.

Pro Tips:

✔ Expand Your Search – Modify the crawler to analyze multiple websites.
✔ Adjust AI Creativity – Lower the temperature for factual reports.
✔ Store in Different Formats – Export results to Google Docs or Notion for better organization.
✔ Enhance the Workflow – Add email notifications when a new thesis is generated.

With this setup, you now have a scalable, AI-powered research and thesis generation system. 

Download the blueprint used in this blog post

Click here to access the blueprint. Whether you’re just starting out or looking to refine your AI Automation skills, our community has got the tools and support to help you succeed.