Edugen AI

Create state-of-the-art AI educational videos, for free

Comment

🚀 Edugen AI — Turning Ideas into Educational Videos

💡 Inspiration

The idea behind Edugen AI came from a simple frustration:

Creating educational videos is slow and costs real money.

Even for short explainers, you usually need:

  • Scripting
  • Designing slides
  • Recording voiceovers
  • Editing timelines

That process can take hours—or even days.

I wanted to explore a different question:

What if generating an educational video felt as simple as typing a topic?

🛠️ How I Built It

Edugen AI is designed as an end-to-end pipeline that converts a topic into a structured video.

At a high level, the system works like this:

  1. Input → Topic
  2. LLM generates structured script
  3. Scene segmentation + timing
  4. Visual generation (slides/frames)
  5. Text-to-speech narration
  6. Final video composition

You can think of it as:

$$ [ \text{Topic} \rightarrow \text{Script} \rightarrow \text{Scenes} \rightarrow \text{Narration} \rightarrow \text{Video} ] $$

The focus wasn’t just automation—but making the output feel coherent and watchable, not robotic.

🧠 What I Learned

Building Edugen AI taught me a few important things:

  • AI output needs structure Raw generation isn’t enough—you need constraints, formatting, and flow.

  • Latency matters more than you think Even a few extra seconds can break the “magic” feeling.

  • Perceived quality > technical complexity Smooth transitions and pacing matter more than fancy internals.

  • UX is everything If the experience isn’t simple, the product feels broken—even if it works.

⚔️ Challenges I Faced

1. Serverless Compute

Very often, the metrics showed spikes in GPU usage. Fix:

  • Parallelised workflow
  • Strict limits.
  • Optimisation in pipelines.

2. Timing & Synchronization: MeDo HELP!

This was one of the hardest parts and MeDo helped me settle this by: Matching:

  • Voiceover
  • Visuals
  • Scene transitions

…was harder than expected.

Fix:

  • Approximate duration modeling
  • Scene-level alignment logic

3. Making It Feel “Alive”

Most AI-generated content feels static.

I wanted:

  • Motion
  • Flow
  • Visual rhythm

Fix:

  • Introduced animation logic
  • Focused on pacing over complexity

4. Speed vs Quality Tradeoff

MeDo Helped me with this as well. $$ [ \text{Better Quality} \uparrow \Rightarrow \text{Latency} \uparrow ] $$

Balancing this was critical.

Goal:

  • Keep generation fast
  • Without making output feel cheap

🔥 What It Does Today

Edugen AI can:

  • Take a topic
  • Generate a structured explanation
  • Convert it into a narrated video
  • Deliver it in less than a minute.

Still early—but already surprisingly usable.

🧠 Final Thought

Edugen AI isn’t just about speed.

It’s about removing friction from explaining ideas.

Because when creating content becomes effortless, more people can focus on what actually matters:

the idea itself

Built With

Share this project:

Updates