An AI-powered visual learning engine that transforms educational text into interactive, code-based 2D/3D visualizations and audio lessons.
It is designed to make learning visual, simple, and accessible for everyone — especially in areas with limited access to high-end tech — by converting text into dynamic visuals and interactive learning experiences using front-end code.
- Fully Live App: https://not-pi-vision.vercel.app/
- GitHub Repo: https://github.com/Sabir-Ali-Mondal/Not-Pi-Vision
Traditional learning methods create significant barriers for modern students:
- Static & Text-Heavy: Study materials are often boring and hard to understand.
- No Visuals: Students struggle with complex/abstract concepts without visual aids.
- Lack of Tools: Teachers and students don't have quick, affordable tools for dynamic study.
- The Digital Divide: Low-resource regions lack access to powerful devices for modern EdTech.
Not Pi Vision uses AI to generate code (HTML/JS/CSS) that renders as live, interactive visuals directly in the browser. It turns any syllabus or topic into a complete visual course.
- Visuals with Voice Narration: Simulates a teacher explaining concepts.
- Book-like Pages with Highlighting: Guides the reader through text content.
- Runs Everywhere: Works on any device with a browser, no powerful hardware needed.
- 🚀 AI Syllabus Deconstruction: Intelligently structures raw text into organized Chapters & Topics.
- 🧠 Multi-Modal Output: Generates interactive Visualizations, audio Podcasts, and readable Pages from a single prompt.
- 🎨 Special Templates (JSON powered): Powers complex, interactive visuals for Chemistry Mechanisms, Graphs, Mindvoice, and Presentations.
- 🔡 Quiz Learning: Quiz Learning adds relevant Quiz as per Learning content.
- 🔗 Smart Image Suggesion: Automatically adds relevant image suggesion of Google Search link to supplement content.
- 🌐 Fully Responsive & Offline: Works on any device and stores data locally for offline access.
- ✨ Special Templates: - JSON powered templates for Fast and Less Faliure and Accurate. - Chemistry Mechanism | Graph | Mindvoice | Presentation
| Feature | 🔍 Not Pi Vision (Code-Based Visuals) | 🧠 Traditional AI Tools (Static Media) |
|---|---|---|
| Visual Generation | Live visuals via code (fast, flexible) | Generates static videos/images (slow) |
| Experience | Dynamic 2D/3D, Interactive, Diagrammable | Often cartoonish or unrealistic |
| Ease of Use | One-click, no heavy setup | Requires editing and re-rendering |
| Performance | Runs on low-end devices, browser-only | Needs high compute & strong internet |
| Technology | Pure frontend (HTML/JS/CSS, p5.js) | GPU-heavy, cloud-dependent |
| Aspect | 🖼️ Not Pi Vision (Code AI) | 🎥 Video Generation AI |
|---|---|---|
| Time to Generate | Instant (1–2 minutes) | High (30 minutes+) |
| Energy/Compute Use | Very Low (text output) | Very High (GPU rendering) |
| Precision | Precise & Diagrammatic for learning | Artistic, often imprecise |
| Interactivity | ✅ Possible | ❌ None (static playback) |
| File Size | Small (few KBs of code) | Large (MBs of video files) |
| Post-Editing | Easy (edit code, instant change) | Hard (requires re-rendering) |
| For Academic Diagrams | ✅ Yes | ❌ No (blurry, unstructured) |
Not Pi Vision uses a cost-effective, two-pronged approach for AI integration:
- Prompt Generation: The app uses a master prompt library (over 1100 lines) to instruct an AI model on how to generate the necessary code and text.
- Frontend Method: It opens ChatGPT/Perplexity in a new tab with the prompt auto-filled. The user copies the AI's output back into the app. No API keys or backend needed.
- Backend Option: For a more seamless flow, higher-end models can be integrated via a Node.js backend.
- Global Adoption: Start with India and similar low-resource regions to maximize impact.
- Continuous Evolution: We are already implementing and developing new features.
- The Ultimate Goal: Enable ChatGPT-like chatbots to provide visualizations directly through frontend-powered code, making interactive visuals accessible to everyone inside a simple chat interface.