Litigence AI is a legal information platform designed to simplify access to Indian legal knowledge through advanced technology. Built with a focus on underserved communities, it provides personalized legal guidance via a conversational AI interface.
Technical Architecture 💻
Frontend: Developed using Flutter for a seamless user experience, currently in closed testing on the Play Console.
Backend: Flask-based backend hosted on Google Cloud Run ensures scalable and efficient processing.
AI Model: Powered by a fine-tuned Vertex AI model with Google's default grounding, enhanced with custom data for improved contextual accuracy.
Cloud Infrastructure: Firebase supports real-time web hosting and future web-based features.
Implemented Features ⭐
Google & OTP Authentication: Simplified and secure user access integrated via Firebase.
Law of the Day: Educates users daily on relevant laws to foster awareness.
Onboarding Flow & Chat Interface: Smooth user onboarding and interactive legal guidance through a conversational chat system.
This application provides a Flask-based API that can be run locally or deployed in a Google Cloud environment. Below are step-by-step instructions on installing dependencies, configuring Google Cloud, running the application, and testing the API locally.
Prerequisites
Python 3 (for running the Flask application)
Google Cloud SDK (for Google Cloud integration)
curl (for API testing)
Running the Application Locally
Clone or open the repository.
Ensure you have installed Python dependencies (if applicable, use pip install -r requirements.txt).
Run the application:
python main.py
This starts the application at http://localhost:8000.
Installing Google Cloud SDK
Use the steps below on a Linux x86_64 machine:
curl -O https://dl.google.com/dl/cloudsdk/channels/rapid/downloads/google-cloud-cli-linux-x86_64.tar.gz
tar -xf google-cloud-cli-linux-x86_64.tar.gz
./google-cloud-sdk/install.sh
(Optional) Add the Google Cloud CLI to your PATH:
# Example approachecho"source ~/google-cloud-sdk/path.bash.inc">>~/.bashrc
source~/.bashrc
Initialize the SDK:
./google-cloud-sdk/bin/gcloud init
Follow the prompts to choose your Google Cloud project and configure…
Testing Magic: Quick generation of comprehensive test cases
Code Refinement: Continuous suggestions for cleaner, better code
GitHub Models 🤖
While Litigence AI currently uses Vertex AI for its production environment (leveraging Google's grounding capabilities and Gemini's custom data integration), we're actively experimenting with GitHub Models, particularly o1, for enhanced RAG capabilities. Our testing involves comparing response accuracy and contextual understanding between different models. The results from these experiments will guide our future model selection, ensuring we deliver the most accurate and reliable legal guidance to our users.
Conclusion 🎯
Building Litigence AI has been an incredible journey for @karthidreamr, transforming a personal mission into a powerful tool for social change. Through the strategic use of GitHub Copilot and cloud technologies, what started as a response to childhood observations of legal inequality is now evolving into a platform that makes legal knowledge accessible to everyone.
The future of Litigence AI looks promising as we continue to enhance features and scale our impact. Together, we're working to ensure that legal awareness becomes a fundamental right, not a privilege! 🌟
1.Chat History: Allow users to revisit previous interactions for better context retention.
2.Web Hosting: Expand accessibility with a web-based platform using Firebase.
3.Error Handling: Implement robust mechanisms like error popups, fallback pages etc for smoother user experience.
4.Test Automation: Write comprehensive test cases to enhance system reliability and performance.
Thank you Github for the student pro plan !
Copilot is awesome, especially the Claude Sonnet 3.5 (new) is a real coder