AI Document Chatbot

## Overview

AI Document Chatbot is a full-stack application that combines Retrieval-Augmented Generation (RAG) with external tool integration to provide intelligent responses from both static documents and dynamic services.

The frontend is built with React and TypeScript, while the backend uses FastAPI and Python to process requests and route them to the appropriate information source.

## Key Features

- Chat interface built with React and Tailwind CSS
- FastAPI backend exposing REST endpoints
- Retrieval-Augmented Generation from PDF, TXT and Markdown files
- Rule-based intent detection for tool routing
- Integration with mocked external services for dynamic data retrieval
- Clean separation between routing, retrieval and tool execution layers

## Architecture

The application routes user questions through an intent detection layer.

- Document-related questions are answered using a lightweight RAG pipeline.
- Dynamic requests (such as remaining vacation days) trigger external service calls instead of document retrieval.

## Design Decisions

The project intentionally avoids external vector databases to remain lightweight and easy to understand while demonstrating clean software architecture and modular design.

## Future Improvements

- Embedding-based semantic search
- LLM-powered intent detection
- Authentication and user-specific data
- Streaming responses
- Docker deployment