Saral Vakeel is a scalable, AI-powered legal assistant hosted on Google Cloud Platform (GCP). It leverages Google Gemini and RAG technology to simplify complex legal documents, offering instant summaries, risk detection, and precise citations with enterprise-grade security and performance.
- 📄 Intelligent Document Analysis: Leverages Google Cloud Document AI for high-accuracy OCR and extracting key metadata including summaries, risk factors, and obligations using Google Gemini 2.5 Flash.
- 💬 RAG-Powered Chat: Ask questions about your uploaded documents and get accurate answers grounded in the text, complete with citations and source highlighting.
- ⚡ Real-Time Processing: Fast PDF ingestion, text chunking, and vector embedding generation.
- ☁️ Enterprise-Grade Scalability: Backend deployed on Google Cloud Platform (GCP), ensuring high availability and the ability to handle massive document workloads effortlessly.
- 🎨 Modern UI/UX: A sleek, accessible interface built with Next.js 15 and Tailwind CSS, featuring dark mode support and interactive background animations.
- 🔍 Precise Citations: Pinpoints the exact page and paragraph where an answer is found.
- 🔒 Secure Storage: Utilizes Supabase for secure file storage and vector database management.
- Framework: Next.js 15 (React 19)
- Styling: Tailwind CSS 4
- Animations: Framer Motion
- Icons: Lucide React
- State/Data: Axios, SWR (implied)
- Framework: FastAPI (Python)
- Cloud Provider: Google Cloud Platform (GCP) (Serverless/Scalable Deployment)
- Document Processing: Google Cloud Document AI (Intelligent OCR & Data Extraction)
- AI/LLM: Google Gemini (Flash 2.5 for Chat & Embedding-004 for Vectors)
- Orchestration: LangChain
- PDF Processing:
pdfplumber
- Vector Store: Supabase (pgvector)
- Authentication & Storage: Supabase Auth & Storage
Before running the project, ensure you have the following installed:
- Node.js (v18 or higher)
- Python (v3.10 or higher)
- Supabase Account (for database and storage)
- Google Cloud API Key (with Gemini API access)
git clone https://github.com/yourusername/saral-vakeel.git
cd saral-vakeelNavigate to the backend directory and set up the Python environment.
cd backend
python -m venv venv
# Activate virtual environment
# Windows:
venv\Scripts\activate
# Mac/Linux:
# source venv/bin/activate
pip install -r requirements.txtCreate a .env file in the root directory (parent of backend) with the following credentials:
SUPABASE_URL=your_supabase_url
SUPABASE_KEY=your_supabase_anon_key
GOOGLE_API_KEY=your_google_gemini_api_keyNavigate to the frontend directory and install dependencies.
cd ../frontend
npm installCreate a .env.local file in the frontend directory:
NEXT_PUBLIC_API_URL=http://localhost:8000
NEXT_PUBLIC_SUPABASE_URL=your_supabase_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_keyFrom the backend directory (with venv activated):
uvicorn main:app --reloadThe backend API will run at http://localhost:8000.
From the frontend directory:
npm run devThe application will be available at http://localhost:3000.
c:\laww\
├── backend/
│ ├── main.py # Main FastAPI application & RAG logic
│ ├── requirements.txt # Python dependencies
│ └── ...
├── frontend/
│ ├── app/ # Next.js App Router pages
│ ├── components/ # React components (Chat, Upload, Visuals)
│ ├── lib/ # Utility functions (Supabase, API)
│ ├── public/ # Static assets
│ └── ...
├── README.md # Project documentation
└── .env # Backend environment variables
Contributions are welcome! Please fork the repository and submit a pull request for any enhancements or bug fixes.
This project is licensed under the MIT License.