Documentabot

An enterprise documentation assistant that lets teams upload PDF documents and ask questions about them via a conversational chat interface. Answers are grounded exclusively in the uploaded content, with citations back to the source file and page number.

---

How it works

1. Upload PDFs to the knowledge base — they are chunked and embedded into a vector store (ChromaDB) using OpenAI embeddings.
2. Ask a question in the chatbot — relevant chunks are retrieved and passed to GPT-4o-mini, which answers using only that context.
3. Each answer includes source citations (file name + page number) so you can verify the response.
4. Chats are persisted per user and can be resumed at any time.

---

Tech stack

Layer	Technology
Frontend	React 19 + TypeScript, Vite, React Router
Backend	FastAPI (Python)
Database	SQLite (users, chats, documents)
Vector store	ChromaDB
Embeddings	OpenAI text-embedding-3-small
LLM	OpenAI gpt-4o-mini
Auth	JWT via HTTP-only cookies

---

Project structure

Documentabot/
├── backend/
│   ├── main.py          # FastAPI app — auth, document, and chat endpoints
│   ├── rag.py           # PDF ingestion, vector retrieval, LLM answering
│   ├── auth.py          # JWT creation, hashing, cookie helpers
│   ├── database.py      # SQLite connection factory
│   ├── migrate.py       # Schema migrations
│   ├── uploads/         # Stored PDF files
│   └── chroma_db/       # Persisted ChromaDB vector store
└── frontend/
    └── src/
        ├── App.tsx              # Routing and nav shell
        ├── components/
        │   ├── Chatbot.tsx      # Chat UI
        │   ├── KnowledgeBase.tsx # Document management
        │   ├── CitationPanel.tsx # Source citation sidebar
        │   ├── UploadModal.tsx  # PDF upload dialog
        │   ├── Login.tsx / Signup.tsx
        │   └── Landing.tsx
        ├── api.ts               # Typed API client
        ├── context/AuthContext.tsx
        └── types.ts

---

Getting started

Prerequisites

• Python 3.11+
• Node.js 20+
• An OpenAI API key

Backend

cd backend
python -m venv venv
venv\Scripts\activate        # Windows
# source venv/bin/activate   # macOS / Linux

pip install fastapi uvicorn python-dotenv langchain langchain-community \
            langchain-openai langchain-chroma chromadb pypdf pyjwt bcrypt

Create a .env file in backend/:

OPENAI_API_KEY=sk-...
SECRET_KEY=your-jwt-secret

Run migrations then start the server:

python migrate.py
uvicorn main:app --reload

The API is available at http://localhost:8000. Interactive docs at http://localhost:8000/docs.

Frontend

cd frontend
npm install
npm run dev

The app is available at http://localhost:5173.

---

API overview

Method	Path	Description
POST	/auth/register	Create an account
POST	/auth/login	Log in
POST	/auth/logout	Log out
GET	/auth/me	Current user
POST	/upload	Upload and index a PDF
GET	/documents	List all documents
DELETE	/documents/{id}	Delete a document and its embeddings
GET	/documents/{id}/view	View PDF inline
GET	/documents/{id}/download	Download PDF
POST	/chats	Start a new chat
GET	/chats	List user's chats
GET	/chats/{id}/messages	Fetch messages for a chat
POST	/chats/{id}/messages	Send a follow-up message