PDFizado

PDFizado is a SaaS I built to make studying with dense PDF material less painful. You upload a document, ask questions in plain Spanish or English, and get answers grounded in the file’s actual content instead of skimming hundreds of pages yourself.

The problem

Students and professionals in Latin America work through long PDFs constantly: textbooks, legal documents, research papers, class notes. Finding the one relevant section means manual scanning, and generic chatbots hallucinate because they never saw the source document.

How it works

The app parses each uploaded PDF, chunks and embeds its content, and retrieves the most relevant passages for every question before the model answers. That retrieval step keeps responses tied to what the document actually says.

• Upload a PDF and start asking questions immediately
• Answers cite the document instead of inventing content
• Extract summaries and key points for faster review
• Subscription plans with usage limits handled through a Colombian payment flow

What I learned

PDFizado was my first end-to-end AI product: prompt design, retrieval quality, token cost control, and billing all in one codebase. Much of my later multi-agent orchestration work started with lessons from keeping this app’s answers reliable.