QueryPilot AI – Multi-User RAG Chatbot
Budget: ₹1,500 – ₹12,500 INR
QueryPilot AI – Multi-User RAG Chatbot
QueryPilot AI is a production-style Retrieval-Augmented Generation (RAG) application that lets users chat with their uploaded documents through a modern React frontend and a FastAPI backend. It combines semantic retrieval, vector search, chat history, and LLM-powered answering to provide context-aware responses from document knowledge.
Overview
This project is built as a multi-user document Q&A system where users can upload files, manage documents, and ask questions in a chat interface. The backend handles retrieval, chat sessions, and response generation, while the frontend provides a clean React-based UI for chatting, uploading, listing, and deleting documents.
Tech Stack
Frontend: React, Vite
Backend: FastAPI
LLM Orchestration: LangChain
Vector Database: ChromaDB
Embeddings / LLMs: Configurable model pipeline
Development Tools: GitHub, VSCode, Linux
Features
Multi-user chat workflow with session-based conversation handling
Document upload, listing, and deletion
Retrieval-Augmented Generation (RAG) pipeline for document-based Q&A
Semantic search over uploaded document content
React frontend with chat UI and collapsible sidebar
FastAPI backend with modular API structure
Easy to extend for production-ready GenAI applications
Architecture
User Query
↓
React Frontend
↓
FastAPI API Layer
↓
Retriever + Vector Search
↓
Relevant Context
↓
LLM Response Generation
↓
Answer returned to chat UI
Project Flow
User uploads one or more documents.
Documents are processed and stored in the vector database.
User asks a question from the React chat interface.
FastAPI sends the query through the RAG pipeline.
Relevant chunks are retrieved from the knowledge base.
The LLM generates a context-aware answer.
The response is shown in the chat window.
Project Structure
querypilot-ai/
│
├── backend/
│ ├── app/
│ ├── main.py
│ ├── schemas.py
│ ├── db_utils.py
│ ├── langchain_utils.py
│ └── requirements.txt
│
├── frontend/
│ ├── src/
│ │ ├── App.jsx
│ │ ├── components/
│ │ │ └── Sidebar.jsx
│ │ └── style.css
│ ├── package.json
│ └── vite.config.js
│
├── chroma_db/
├── RAG_Docs/
├── .gitignore
└── README.md
Current Frontend Capabilities
Model selection
New chat reset
File upload
Uploaded document list
Delete selected document
Chat interface for document-based Q&A
Collapsible left sidebar layout
Future Improvements
General chat mode in addition to document-only RAG mode
Authentication for real multi-user access
Better chat history management
Streaming responses
Deployment with Docker and cloud hosting
Getting Started
Backend
cd backend
pip install -r requirements.txt
uvicorn main:app --reload
Frontend
cd frontend
npm install
npm run dev
Use Case
This project is useful for building intelligent assistants over PDFs, reports, notes, and internal documents. It can be extended into enterprise search, knowledge assistants, customer support systems, and internal AI copilots.
Author
Built as a hands-on GenAI engineering project focused on RAG, FastAPI, React, and production-oriented system design.
QueryPilot AI is a production-style Retrieval-Augmented Generation (RAG) application that lets users chat with their uploaded documents through a modern React frontend and a FastAPI backend. It combines semantic retrieval, vector search, chat history, and LLM-powered answering to provide context-aware responses from document knowledge.
Overview
This project is built as a multi-user document Q&A system where users can upload files, manage documents, and ask questions in a chat interface. The backend handles retrieval, chat sessions, and response generation, while the frontend provides a clean React-based UI for chatting, uploading, listing, and deleting documents.
Tech Stack
Frontend: React, Vite
Backend: FastAPI
LLM Orchestration: LangChain
Vector Database: ChromaDB
Embeddings / LLMs: Configurable model pipeline
Development Tools: GitHub, VSCode, Linux
Features
Multi-user chat workflow with session-based conversation handling
Document upload, listing, and deletion
Retrieval-Augmented Generation (RAG) pipeline for document-based Q&A
Semantic search over uploaded document content
React frontend with chat UI and collapsible sidebar
FastAPI backend with modular API structure
Easy to extend for production-ready GenAI applications
Architecture
User Query
↓
React Frontend
↓
FastAPI API Layer
↓
Retriever + Vector Search
↓
Relevant Context
↓
LLM Response Generation
↓
Answer returned to chat UI
Project Flow
User uploads one or more documents.
Documents are processed and stored in the vector database.
User asks a question from the React chat interface.
FastAPI sends the query through the RAG pipeline.
Relevant chunks are retrieved from the knowledge base.
The LLM generates a context-aware answer.
The response is shown in the chat window.
Project Structure
querypilot-ai/
│
├── backend/
│ ├── app/
│ ├── main.py
│ ├── schemas.py
│ ├── db_utils.py
│ ├── langchain_utils.py
│ └── requirements.txt
│
├── frontend/
│ ├── src/
│ │ ├── App.jsx
│ │ ├── components/
│ │ │ └── Sidebar.jsx
│ │ └── style.css
│ ├── package.json
│ └── vite.config.js
│
├── chroma_db/
├── RAG_Docs/
├── .gitignore
└── README.md
Current Frontend Capabilities
Model selection
New chat reset
File upload
Uploaded document list
Delete selected document
Chat interface for document-based Q&A
Collapsible left sidebar layout
Future Improvements
General chat mode in addition to document-only RAG mode
Authentication for real multi-user access
Better chat history management
Streaming responses
Deployment with Docker and cloud hosting
Getting Started
Backend
cd backend
pip install -r requirements.txt
uvicorn main:app --reload
Frontend
cd frontend
npm install
npm run dev
Use Case
This project is useful for building intelligent assistants over PDFs, reports, notes, and internal documents. It can be extended into enterprise search, knowledge assistants, customer support systems, and internal AI copilots.
Author
Built as a hands-on GenAI engineering project focused on RAG, FastAPI, React, and production-oriented system design.