GenAI Legal Assistant Chatbot

A powerful, modular legal chatbot that leverages multi-agent architecture to parse, retrieve, and summarize complex Indian legal documents in plain English — all through an interactive Streamlit interface. - Test here

Supports multiple LLM providers:

• OpenAI (ChatGPT)
• Groq (LLM-as-a-service)
• Gemini (Google)
• Ollama (local models like LLaMA3) - Only for local usage

---

Features

• PDF Ingestion & Chunking with FAISS vector search
• Query-based Retrieval from embedded legal content
• Summarization Agent using the selected LLM
• Context-aware Chat Looping
• Save/Load Full Session (incl. history + model + API keys)
• Custom API key input per provider
• Local file dialogs for session saving
• Clear-all button resets everything cleanly

---

System Architecture

---

Tech Stack

Component	Tool/Library
UI	Streamlit
Vector DB	FAISS
Embeddings	Sentence Transformers
PDF Parser	PyMuPDF
LLM Support	OpenAI, Gemini, Groq, Ollama
File Dialog	Tkinter

---

Project Structure

genai-legal-bot
├── app.py                   # Main Streamlit UI app
├── pyproject.toml           # Poetry configuration
├── saved_sessions/          # Folder for saved JSON chat sessions
├── knowledge_base/          # Legal PDFs loaded into the system
│   ├── Guide_to_Litigation_in_India.pdf
│   └── Legal_Compliance_ICAI.pdf
├── data/
│   └── chunks.json          # Generated PDF chunks
├── agents/
│   ├── query_engine.py      # Handles vector search + retrieval
│   └── summarizer_agent.py  # Summarizes chunks using selected model
└── README.md

---

Setup Instructions

1. Clone the Repo

git clone https://github.com/yourusername/genai-legal-bot.git
cd genai-legal-bot

2. Install Poetry & Dependencies

pip install poetry
poetry install

3. Add PDF files to knowledge_base/

Put any PDF documents you want to embed into the chatbot.

4. Run the App

poetry run streamlit run app.py

---

Supported Model Providers & How to Use

OpenAI

• Paste your OPENAI_API_KEY in the sidebar.
• Choose a model like gpt-3.5-turbo.

Gemini

• Paste your GOOGLE_API_KEY.
• Models like gemini-1.5-pro-latest will appear.

Groq

• Paste your GROQ_API_KEY.
• Select from llama3-8b, mixtral, or others.

Ollama (local)

• Requires Ollama installed locally
• Make sure models like llama3 are available via ollama list

---

Example Use Flow

1. Upload your legal PDFs

2. Start typing a legal query like:

   "What is the process to file a summary suit in India?"

3. Choose a model & enter API key (if needed)

4. View the retrieved legal sections + a plain-English summary

5. Save the session to pick up later

---

Commands & Dev Utilities

Clear Chunks & Re-Embed:

Delete data/chunks.json and re-run the app to regenerate chunks.

---

License

MIT License — use, modify, and build freely.

---

For support or contributions, open a GitHub issue or email: chagantivenkataramireddy1@gmail.com