An offline-first AI math solver built for Indian students. It doesn't just give answers — it walks you through the solution the way a teacher would, following the "Given, Formula, Solution, Result" structure used in CBSE, ICSE, and State Board classrooms.
NeuraMath takes a math problem in plain English or standard notation and returns a step-by-step breakdown with clear reasoning at each stage. Everything runs locally in the browser with no server calls or API keys required.
Examples of what you can ask:
- "Solve 2x + 5 = 15"
- "Find the area of a circle with radius 7"
- "Differentiate x^3 + 2x"
- "Convert 5 km to miles"
- Class 6-8: Arithmetic, fractions, percentages, simple equations
- Class 9-10: Quadratics, trigonometry, statistics, surface area and volume
- Class 11-12: Calculus, matrices, complex numbers
- Linear and quadratic equations with discriminant analysis
- Derivatives and integrals (symbolic, via nerdamer)
- Geometry calculations for circles, triangles, rectangles, spheres, and cubes
- Statistics: mean, median, mode, variance, standard deviation
- Trigonometric functions with degree/radian handling
- Unit conversions (km/miles, kg/lbs, Celsius/Fahrenheit, rupees/paise)
- Number theory and word problem parsing
- Dark glassmorphic UI with a light theme toggle
- KaTeX rendering for textbook-quality math notation
- Collapsible step-by-step explanations
- Canvas-based graph plotting for quadratic functions
- Voice input via Web Speech API
- Keyboard shortcuts (Enter to solve, Escape to clear)
- Problem history stored locally
- Practice problem generator
- PDF export and solution sharing
- Progress tracking and mastery dashboard
- Fully offline — no backend, no API dependencies
- Built with vanilla HTML, CSS, and JavaScript
- Uses KaTeX, nerdamer, and math.js via CDN
- PWA-ready with service worker and manifest
- localStorage for persistence and learning
- Clone the repository:
git clone https://github.com/ayanhackss/neuramath.git
- Open
index.htmlin any modern browser.
That's it. No build step, no dependencies to install.
index.html Main application page
app.js UI controller and interaction logic
neura-engine.js AI classification engine and solvers
index.css Styling (dark/light themes, glassmorphism)
sw.js Service worker for offline support
manifest.json PWA manifest
prd.md Product requirements document
- You type or speak a math problem.
- The neural classification engine identifies the problem type using keyword pattern matching and confidence scoring.
- The appropriate solver processes the input symbolically.
- Results are displayed with step-by-step reasoning, rendered in proper math notation.
All computation happens in the browser. The engine learns from usage patterns via localStorage to improve classification over time.
- HTML5, CSS3, JavaScript (ES6+)
- KaTeX — math rendering
- nerdamer — symbolic algebra and calculus
- math.js — numerical evaluation
- Web Speech API — voice input
- Canvas API — graph plotting
Contributions are welcome. If you'd like to add a new solver, improve classification accuracy, or fix a bug:
- Fork the repository
- Create a feature branch
- Make your changes
- Open a pull request with a clear description of what you changed and why
This project is open source. See the LICENSE file for details.
Built as a learning tool for Indian students who deserve a math tutor that explains things the way their teachers do — step by step, in language they understand.