DEV Community

Cover image for How I Built a Tool to Turn PDFs into Videos
Hazel
Hazel

Posted on

How I Built a Tool to Turn PDFs into Videos

I didn’t plan to build a product at first. It started with a stack of PDFs on my desk and a thought:
“Why does all this content feel so dead?”

I love reading research papers, guides, and articles, but PDFs are static. Sharing them with friends or teammates? Even worse—nobody wants to open a 50-page document. So one weekend, I decided to hack something together.

Stage 1: Plain Text on a Black Screen

The first version was embarrassingly simple: black background, white text, auto-scrolling.

I had a simple text renderer with a timer to scroll lines. No styling, no animations. It worked… but it was boring.

Stage 1: Plain Text on a Black Screen

Stage 2: Adding Background Images

Next, I thought: “This needs some life.” I added background images. Suddenly, it felt more like a presentation than a debug tool.

Overlaying text on images while keeping it readable was tricky. I added a layering system with text wrapping and contrast checks, which made the output much cleaner.

Adding Background Images

Stage 3: Adding Voice

Then came the real turning point—voice narration. Reading text out loud made the content engaging and easy to follow.

Syncing the narration with the scrolling text was challenging. At first, the voice would finish before the text or lag behind. After some tweaks, I managed to align them smoothly, and the video finally felt natural.

Stage 3: Adding Voice

Stage 4: Multimodal Inputs

People wanted more than text. PDFs, article links… so I added:

PDFs (up to 80 pages)

Prompts (up to 2000 characters)

Article links

Parsing PDFs was tricky because of different fonts, layouts, and embedded images. I added a normalization step to handle that. Article links required cleaning HTML and extracting the main content. Prompts were simpler but capped to avoid errors.

From a simple text scroller, it evolved into a flexible video generator.

Stage 4: Multimodal Inputs

Why It Feels Rewarding

Every stage came with frustration—PDF parsing, syncing voice, simplifying the UI—but seeing the output improve brought relief and excitement.
“This might actually help someone besides me.”

What the Tool Can Do Today

It’s called PDF to Video
You can:
Upload a PDF, enter a prompt, or paste an article link
Set video length (0–15 minutes)
Choose aspect ratio (1:1, 16:9, 9:16)
Pick a narration voice
Customize theme, title, and even upload a logo

The core idea remains: turning static documents into videos that are alive, watchable, and shareable.

What the Tool Can Do Today

Looking Ahead

Looking back at that first black screen, I almost laugh—but I’m grateful. Without that ugly start, I wouldn’t have pushed forward.

Hearing users say: “It saved me hours of work” or “I can study better now”—that’s the most rewarding part.

Building this was messy and exhausting, but it’s also the project I’m proudest of.

And next time you stare at a boring PDF, think: “What if this could be a video?”

👉 That’s exactly what PDF to Video was born for.
https://pdftovideo.ai/

Top comments (0)