Document AI Agent

A Streamlit chatbot that answers questions from .docx documents using Azure OpenAI.

Drop Word documents into the docs/ folder and ask questions in a conversational chat interface. The AI answers strictly from your documents and maintains chat history across turns.

How It Works

1. loader.py reads all .docx files from docs/, extracting text with section headings preserved
2. agent.py sends the full document text, conversation history, and your question to Azure OpenAI
3. app.py provides the Streamlit chat UI with message history and a clear-chat button

Setup

Prerequisites

• Python 3.12+
• An Azure OpenAI deployment (e.g., gpt-4.1-mini)

Install

python -m venv .venv
.venv\Scripts\Activate.ps1  # Windows
pip install -r requirements.txt

Configure

Create .streamlit/secrets.toml with your Azure OpenAI credentials:

AZURE_OPENAI_API_KEY = "your-azure-api-key"
AZURE_OPENAI_ENDPOINT = "https://your-resource.openai.azure.com/"
AZURE_OPENAI_DEPLOYMENT = "your-deployment-name"
AZURE_OPENAI_API_VERSION = "2025-01-01-preview"

Run

streamlit run app.py

Usage

1. Place .docx files in the docs/ folder
2. Start the app and ask questions in the chat input
3. The AI will answer based only on the document contents
4. Use the clear chat button to reset the conversation

Project Structure

app.py                  # Streamlit chat interface
agent.py                # Azure OpenAI integration
loader.py               # Word document text extraction
docs/                   # Drop .docx files here
.streamlit/secrets.toml # Azure OpenAI credentials (not tracked in git)
requirements.txt        # Python dependencies