StudyReels
Inspiration
With attention spans shrinking and short-form content dominating the internet, traditional learning methods often fail to keep students engaged. I wanted to make studying as entertaining and effortless as watching a reel. Inspired by platforms like YouTube Shorts and Duolingo, StudyReels was born — an AI-powered tool that transforms educational content into short, dialogue-driven videos that feel more like stories than lectures.
What it does
StudyReels turns any topic into an engaging, dialogue-based explainer video.
- The user enters a topic (e.g., Quantum Physics or Photosynthesis).
- The AI generates a script between two characters discussing the topic conversationally.
- Realistic voiceovers are created using Hugging Face’s nari-labs/Dia-1.6B model.
- FFmpeg then syncs the voiceovers, subtitles, and gameplay visuals into a short video.
The result is a fun, bite-sized educational reel — powered entirely by AI.
How we built it
- Frontend: Built with React + TypeScript, offering an intuitive interface to enter topics and preview generated reels.
- Backend: Developed using Node.js, handling API requests, AI script generation, and media synthesis.
- Script Generation: Implemented using the Gemini API, which produces natural dialogues between two characters.
- Voice Synthesis: Integrated Hugging Face’s TTS model to generate lifelike speech.
- Video Assembly: Utilized FFmpeg for merging video, subtitles, and audio seamlessly.
Challenges we ran into
- Synchronizing TTS audio timing with subtitle and video playback.
- Managing latency from large model APIs during real-time generation.
- Ensuring consistency in tone and clarity between multiple AI-generated voices.
- Optimizing FFmpeg processing speed without compromising video quality.
Accomplishments that we're proud of
- Successfully automated the complete video creation pipeline — from text input to ready-to-watch educational reels.
- Built an AI workflow that blends scriptwriting, speech synthesis, and video generation into one seamless process.
- Created a product that makes studying genuinely enjoyable and accessible to a global audience.
What we learned
- Deep understanding of AI content generation, TTS models, and media automation.
- Improved skills in prompt engineering, asynchronous backend design, and pipeline optimization.
- Learned how to design a product that balances technical complexity with user experience.
What's next for StudyReels
- Introducing personalized voices and avatars for characters.
- Adding multilingual support for global accessibility.
- Integrating interactive quizzes within videos to enhance learning retention.
- Expanding the library of visuals and styles to make every reel unique.
Built With
- express.js
- ffmpeg
- gemini-api
- github
- hugging-face-(nari-labs/dia-1.6b)
- mongodb
- node.js
- railway
- react
- render
- typescript
- vercel
Log in or sign up for Devpost to join the conversation.