Neo AI Tutor

This AI tool turns any math problem into a clear, animated video with voice narration—making math easier, visual, and more engaging for all learners, anywhere.

Comment

🌟 Inspiration

We’ve always admired how the YouTube channel 3Blue1Brown makes complex mathematical ideas feel intuitive—and even beautiful—through Manim animations.
That got us thinking:
What if anyone could generate videos like that automatically?
We wanted to bring that same level of clarity and magic to learners everywhere—especially those without access to high-quality math teachers or visual resources.

What It Does

Our platform takes any math problem—whether typed or handwritten—and turns it into a step-by-step animated video with synchronized voice narration.
It’s like having a personal AI math tutor who explains concepts visually and verbally—making learning more engaging, clear, and memorable.

How We Built It

  • Manim – to generate clean, structured math animations
  • OpenAI / Gemini – to convert math problems into Manim-ready scripts
  • ElevenLabs / gTTS – to produce natural, clear audio explanations
  • MoviePy – to combine animations and narration into a final video
  • Streamlit – for building an accessible and responsive web interface
  • Tesseract OCR – to extract math content from handwritten images

Challenges We Ran Into

  • Generating error-free Manim scripts using LLMs required extensive prompt engineering and validation.
  • Synchronizing the timing between visuals and audio using MoviePy was more complex than expected.
  • Ensuring that animations were clean, non-overlapping, and easy to understand took iterative tweaking.

Accomplishments We're Proud Of

  • Built a fully automated system that converts math problems into animated, narrated educational videos.
  • Added support for handwritten input, enabling users to upload photos of math problems.
  • Created a solution that can genuinely improve math education for students by combining AI, animation, and audio.

What We Learned

  • Manim is powerful—but unforgiving. Automating it taught us how to debug and validate visual code at scale.
  • Prompt engineering is key—especially when combining multiple AI models across different tasks.
  • Visual + audio learning dramatically boosts clarity and engagement for students.

What's Next

  • Add multi-language and accent support for voiceovers to reach more global learners
  • Expand coverage to topics like geometry, trigonometry, and competitive exams
  • Launch a full teacher portal where educators can generate and share custom videos
  • Add interactive quizzes to help students watch, learn, and practice—all in one platform

Built With

Share this project:

Updates