I've reviewed every PDF editor out there - then I had ChatGPT build me a better one

At a glance:

• ChatGPT generated a Python script to remove yellow backgrounds from PDFs, solving a real-world problem
• The solution leverages AI's non-deterministic nature to create a deterministic tool
• The script handles multi-page PDFs and preserves text quality

The Problem

Denise, a church choir member, needed to digitize yellow-backed sheet music for practice. Traditional methods like Photoshop were time-consuming, requiring manual adjustments for each image. She sought a faster, more reliable solution to convert the music into clean black-and-white PDFs without losing detail. The challenge was twofold: removing the yellow background while maintaining readability and ensuring the output matched the original sheet music exactly. Her concern about AI altering the content added another layer of complexity, as non-deterministic AI tools could produce inconsistent results.

The AI Solution

Instead of relying on manual editing, I turned to ChatGPT. Despite its reputation for variable outputs, I crafted a precise prompt asking for a Python script that would process PDFs by converting non-gray/black pixels to white. The script needed to handle multi-page documents and preserve text integrity. ChatGPT delivered a functional solution within minutes, bypassing the need for complex UI design or iterative testing. This approach highlighted a key insight: AI can be used not just for direct tasks but to generate deterministic tools. Denise's use case demonstrated how AI can solve niche problems when paired with clear, algorithmic specifications.

The Python Script

The resulting script, decolor_pdf.py, operates via command-line arguments. It analyzes each page of a PDF, identifies pixels outside the gray/black spectrum, and replaces them with white. This process ensures color printers aren't wasted on yellow backgrounds while keeping text sharp. Testing confirmed it works flawlessly on single and multi-page PDFs. The script's simplicity is its strength—no graphical interface, just a text-based tool that runs in seconds. I shared the code on GitHub, making it accessible for others facing similar challenges. The script's success underscores Python's versatility in automating tasks once thought to require manual intervention.

How It Works

The script uses Python libraries like PyMuPDF to parse PDFs and manipulate images. It scans each page pixel by pixel, applying a threshold to distinguish yellow from gray/black. Pixels that don't meet the threshold are whitened. This method is efficient because it avoids complex image processing algorithms, relying instead on straightforward color analysis. The deterministic nature of the script—despite ChatGPT's non-deterministic origins—ensures consistent results every time. This contrast between AI's unpredictability and the script's reliability illustrates a unique use case: using AI to create tools that are inherently deterministic.

Lessons Learned

This experience taught several valuable lessons. First, AI doesn't always need to be used directly; sometimes, it's better to ask it to generate code or tools. Second, command-line tools can be more practical than graphical interfaces for specific tasks. Third, refining AI prompts iteratively can yield precise outcomes. Denise's case also shows how AI can address accessibility issues, like making sheet music easier to read. Looking ahead, such tools could expand to other domains, from document editing to data cleaning, proving that AI's potential extends beyond its most obvious applications.

Future Implications

While the script solves a specific problem, it raises questions about AI's role in automation. As AI tools become more sophisticated, they may increasingly generate custom solutions for niche needs. However, users must remain cautious about non-deterministic outputs. In Denise's case, the script's reliability was critical, but in other scenarios, variability could be a drawback. The balance between AI's flexibility and the need for precision will likely shape its adoption in professional and personal workflows. Additionally, this project highlights the importance of open-source solutions, as sharing the script allows others to benefit without reinventing the wheel.