VeriWrite is an innovative, full-stack web-based platform engineered to provide a robust solution for detecting plagiarism and ensuring content originality. While initially focused on academic submissions, its core capabilities extend to general document verification. It leverages a sophisticated blend of text extraction, advanced similarity detection algorithms, and real-time web content comparison to generate insightful, detailed plagiarism reports. VeriWrite aims to empower educators and content creators, preserve academic integrity, and streamline the evaluation process.
- Comprehensive Document Input:
- Supports multi-format uploads including PDFs, Word Documents (DOCX/DOC), and various image files (JPG, JPEG, PNG).
- Intelligent Text Extraction: Employs Microsoft Azure OCR API for highly accurate text extraction from image-based content and scanned documents (including handwritten text). Features a robust fallback mechanism for PDFs: if standard PDF parsing yields no or minimal text, it automatically switches to OCR to ensure comprehensive text capture.
- High-Performance Plagiarism Engine:
- Optimized Comparison: Architected a plagiarism core that achieves a remarkable 95% reduction in pairwise comparison overhead by implementing advanced techniques like MinHash and Locality Sensitive Hashing (LSH). This ensures rapid and scalable analysis, even with large datasets.
- Accurate Similarity Scoring: Utilizes Jaccard-Overlap token scoring for precise measurement of content similarity.
- Real-time Online Plagiarism Check:
- Integrates with SERP API to compare uploaded content against billions of live web pages, providing dynamic and comprehensive online matching.
- Detailed & Actionable Reports:
- Auto-generates grouped plagiarism reports for submissions with high similarity.
- Reports include highlighted matches, a clear similarity score breakdown, and options for in-browser viewing and PDF download.
- Secure Virtual Classroom System:
- Features role-based access control (RBAC) for teachers and students.
- Teachers can create assignments and exams, set deadlines, and track submissions.
- Students can upload their assignments, with late submission handling.
- All submissions are seamlessly mapped to specific classes and user IDs.
- Secure Authentication: Robust user authentication system supporting both Email/Password.
- Intuitive User Dashboards: Dedicated dashboards for both teachers and students, providing organized views of courses, assignments, submissions, and plagiarism results.
| Layer | Technologies Used |
|---|---|
| Frontend | React.js, Tailwind CSS (for modern UI/UX and custom animations), HTML/CSS (UI components) |
| Backend | Node.js, Express.js (RESTful API development) |
| Database | MongoDB (via Mongoose ORM) |
| Algorithms | MinHash, Locality Sensitive Hashing (LSH), Jaccard-Overlap, N-grams/Shingles |
| APIs/Services | Microsoft Azure OCR API, SERP API |
| Authentication | JSON Web Tokens (JWT) |
| File Handling | Multer, PDF-Parse, Mammoth.js, Node's fs module |
| Deployment | Vercel (for Frontend), Render (for Backend) |
Experience VeriWrite live and see its core functionalities in action!
🔗 Live Demo: VeriWrite
An overview of the intuitive dashboard for both teachers and students, providing quick access to courses and assignments.
A look at how assignments are presented and the streamlined process for student submissions.
An example of the comprehensive plagiarism report, highlighting matched sections and similarity scores.
A clean interface for uploading documents and initiating plagiarism checks against online sources.
To get VeriWrite up and running on your local machine, follow these steps:
# 1. Clone the repository
git clone https://github.com/Sid-CodeX/VeriWrite.git
cd VeriWrite
# 2. Install Frontend Dependencies & Start
cd frontend
npm install
npm run dev
# 3. Install Backend Dependencies & Start
cd ../backend
npm install
npm run dev 🔑 Environment Variables (.env files):
Ensure you create .env files in both your frontend and backend directories. Refer to the .env.example (or similar) files within those folders for required variables such as:
MONGODB_URIJWT_SECRETAZURE_OCR_ENDPOINT,AZURE_OCR_KEYSERP_API_KEYVITE_API_BASE_URL(for frontend)
- Student Upload: A student uploads an assignment (PDF, DOCX, image).
- Text Extraction:
- For PDFs,
pdf-parseattempts extraction. If minimal/no text, Azure OCR is used as a fallback. - For Word Docs,
mammoth.jsextracts raw text. - For images, Azure OCR extracts text.
- For PDFs,
- Preprocessing & Shingling: The extracted text is normalized, tokenized, and converted into 'shingles' (fixed-size word sequences).
- MinHash Signature Generation: Each document's shingles are used to generate a compact MinHash signature.
- Locality Sensitive Hashing (LSH): Signatures are partitioned into bands and hashed, allowing for efficient candidate pair identification (documents likely to be similar) without N^2 comparisons. This is where the 95% overhead reduction is achieved.
- Jaccard Similarity Calculation: For candidate pairs, the Jaccard similarity of their original shingle sets (or MinHash signatures) is calculated to determine the precise degree of overlap.
- Online Comparison (SERP API): For the "Online Plagiarism Check," extracted text segments are queried against web content via the SERP API.
- Report Generation: A detailed PDF report is compiled, grouping similar submissions and highlighting plagiarized sections.
- Automated Assignment/Exam Evaluation: Integrate AI/ML models for automatic grading of exams or assignments based on teacher's model answers, extending beyond just plagiarism detection.
- Semantic Plagiarism Detection: Implement advanced NLP techniques (e.g., using BERT embeddings) to detect conceptual or rephrased plagiarism, rather than just exact string matches.
- Fine-tuned OCR: Integrate and fine-tune models like Tesseract for even more precise text extraction, particularly for diverse handwritten styles or specialized document types.
- AI-driven Feedback: Provide AI-generated constructive feedback on student submissions.
Proudly developed as part of a 6th-semester Mini Project at Muthoot Institute of Technology and Science (MITS), under the esteemed guidance of Dr. Sreenu G, CSE Department.
This project is licensed under the MIT License.
We welcome contributions! Feel free to fork this repository, suggest enhancements via issues, or submit pull requests. For major changes, please open an issue first to discuss what you would like to change.
Have feedback or questions? Reach out via GitHub Issues or connect with any team member on LinkedIn.
Engineered with precision and passion by Team VeriWrite