NeuroClass is an AI-enhanced Learning Management System (LMS) that brings the familiar, streamlined evaluation workflow of Google Classroom into the age of artificial intelligence. It allows instructors to create assignments and students to submit work, but replaces traditional manual grading with a high-speed, forensic AI evaluation engine.
By combining the intuitive assignment management of Google Classroom with the analytical power of Groq's Llama-3 models, NeuroClass ensures students receive instant, high-quality feedback while instructors maintain complete control through custom rubrics and "gold standard" training samples.
Frontend (React + Vite)
↓
Backend API (Node.js + Express)
↓
AI Engine (Groq + Gemini)
↓
Vector DB + PostgreSQL (Supabase)
NeuroClass mimics the intuitive flow of Google Classroom to make it instantly familiar:
- Assignment Creation: Instructors post assignments, attach rubrics, and provide baseline "Gold Samples."
- Student Submission: Students submit answers via text, PDF, or DOCX, just like in a standard LMS.
- AI Evaluation: The AI instantly grades the work based on the instructor's rubric (replacing manual review).
- Instant Results: Students get their marks and feedback immediately in their dashboard.
- Leaderboard Update: Successful submissions are automatically ranked on the course leaderboard.
Powered by Groq (Llama 3.3 70B), our evaluation engine goes beyond simple grading.
- Relevancy Gate: Automatically detects off-topic or "Hallucinated" answers and applies a zero-tolerance policy.
- Gap Analysis: Identifies exactly what concepts are missing from a student's answer based on the instructor's rubric.
- Evidence-Based Grades: Forced to quote the student's work to justify every point awarded or deducted.
Built-in peer-to-peer similarity checks.
- Calculates Cosine Similarity between submissions within the same assignment.
- Flags suspicious overlaps to instructors before final scores are published.
A dedicated chat assistant for students.
- Non-Directive Support: Guides students toward the answer using hints and logic instead of giving away solutions.
- Course Context: Uses course-specific materials (PDFs, uploads) to provide relevant guidance.
- Live Rankings: Automatic student spotlight for Top 3 performers (Gold, Silver, Bronze badges).
- Course Specificity: Filter rankings by subject or view global performance.
Instructors can fine-tune the AI by:
- Uploading custom Rubrics.
- Providing Gold Standard Samples (High/Medium/Low quality) for few-shot learning accuracy.
| Layer | Technology |
|---|---|
| Frontend | React (Vite), Tailwind CSS, Lucide React, Firebase Auth |
| Backend | Node.js, Express, Multer, PDF-Parse, Mammoth |
| Database | Supabase (PostgreSQL) |
| AI Engine | Groq (Llama 3.1 & 3.3), Google Gemini |
| Styling | Vanilla CSS (Modern Aesthetics), Glassmorphism |
git clone https://github.com/DarshanCodes09/neuroclass.git
cd neuroclassCreate a .env file in both client and server folders.
Server (.env):
PORT=8000
SUPABASE_URL=your_supabase_url
SUPABASE_SERVICE_ROLE_KEY=your_service_role_key
GROQ_API_KEY=your_groq_keyClient (.env):
VITE_SUPABASE_URL=your_supabase_url
VITE_SUPABASE_ANON_KEY=your_anon_key
VITE_AI_BACKEND_URL=http://localhost:8000/api# From the root directory
npm run install:all# Start Backend (Port 8000)
npm run dev:server
# Start Frontend (Vite)
npm run dev:clientWe welcome contributions from the community! To contribute to NeuroClass:
- Fork the repository.
- Create a new branch (
git checkout -b feature/amazing-feature). - Commit your changes (
git commit -m 'Add some amazing feature'). - Push to the branch (
git push origin feature/amazing-feature). - Open a Pull Request.
Distributed under the MIT License. See LICENSE for more information.
Developed with ❤️ by Darshan Kushalkar.