VibeStudy

Imagine learning anything you want from a short video instantly. Our project transforms any text prompt or image related to science and mathematics into AI generated video with visualization.

Comment

Inspiration

We were inspired by the way short videos have transformed how people learn online — from TikTok explainers to YouTube tutorials. However, creating a high-quality explainer video still takes a lot of time, skill, and effort. We asked ourselves: what if you could simply type a prompt and get a full video generated for you — instantly? That's how the idea for Instant Video Explainer was born: make learning and content creation as easy as writing a sentence.

What We Learned

Throughout the hackathon, we learned:

  • How to integrate AI models for video, audio, and script generation.
  • Techniques for synchronizing voiceover narration with AI-generated visuals.
  • Best practices for fast, lightweight API communication to keep generation times short.
  • How to balance creative control with full automation in user experience.
  • How We Built It

Our stack included:

  • Backend: Python (Flask) to orchestrate the workflow.
  • AI Models: LLM (like GPT) for scriptwriting based on the user's prompt.
  • Text-to-Speech (TTS) model for generating natural-sounding narration.
  • Frontend: A minimal web interface using React where users enter prompts and receive downloadable videos.

Workflow:

  1. User enters a prompt.
  2. Backend generates a short script explaining the concept.
  3. Script is fed into TTS to create the audio narration.
  4. Visuals are generated or pulled from AI image/video models.
  5. Final video is stitched together and delivered to the user.

Challenges We Faced

  1. Synchronization between narration and visuals was tricky, ensuring that visuals matched the flow of the audio required fine-tuning timing algorithms.
  2. Speed: we had to optimize processing to make sure video generation felt "instant" without overwhelming our servers.
  3. Content Quality: balancing the detail and simplicity of generated scripts to make sure videos were actually educational and not just generic.
  4. Limited Resources: high-quality video generation is compute-intensive; we had to creatively use lightweight models and techniques within our hackathon's time and resource constraints.

Built With

Share this project:

Updates