Juris-Cape: AI Legal Assistant

Juris-Cape is a multimodal AI legal assistant that uses Google Gemini to analyze crime scene photos, surveillance videos, and legal documents. It answers legal questions by cross-referencing valid laws (like the Bharatiya Nyaya Sanhita - BNS).

---

🚀 How to Run (Step-by-Step)

Follow these instructions to set up the project on your Windows machine.

Step 1: Prerequisites

• Install Python 3.10 or higher (Download Here).
• Get a Google Gemini API Key (Get it for free here).

Step 2: Configure Environment

1. Create a .env file:
   • Look for a file named .env in the project folder.
   • If it doesn't exist, create a new text file named .env.
   • Open it and add your key like this:

     GEMINI_API_KEY=your_actual_api_key_here

   • Save the file.

Step 3: Install Dependencies

Open your terminal (PowerShell or Command Prompt) in the juris- cape folder and run:

pip install -r requirements.txt

(If you are using a virtual environment, make sure it is activated or use the full path to python as shown in recent attempts).

Step 4: Verify Everything Works

Before running the server, run the verification script to test Image, Document, and Video analysis.

python verify_multimodal.py

• It will verify your API key is working.
• It will test image analysis on a generated red square.
• You can optionally paste a path to a video file to test video analysis.

Step 5: Start the App

Run the backend server:

python -m uvicorn app.main:app --reload

You should see output saying Application startup complete.

Step 6: Use the App

Open your browser and verify the API at: 👉 http://127.0.0.1:8000/docs

You can use the Try it out button on endpoints like /gemini/chat to ask legal questions.

---

📂 Project Overview

• app/services/gemini_service.py: The brain of the operation. Handles all interactions with Google Gemini.
• app/api/endpoints/gemini_routes.py: The API links used by the web interface.
• data/vector_db: Detailed storage for legal precedents (RAG).

🆘 Troubleshooting

• "Module not found" error? Try running with python -m pip install -r requirements.txt again.
• "Quota exceeded" error? We are using the free tier of Gemini. If you hit limits, wait a minute and try again.
• "Unicode/Emoji" errors? Your terminal might not support emojis. The script is designed to handle this, but if it crashes, try using a modern terminal like Windows Terminal.