A simple browser-based repository assistant built with Python and Azure AI. The app serves a static web UI from static/ and routes chat prompts through server.py to an Azure AI project agent.
- Serves a lightweight HTML/CSS/JavaScript chat interface
- Sends user questions to an Azure AI project agent
- Displays the agent response in a modern chat workflow
- Supports repository question answering, code explanation, and documentation summary
server.py- Python HTTP server that forwards user prompts to Azure AIstatic/index.html- Web UI shellstatic/style.css- Visual styling for the chat appstatic/script.js- Front-end logic for sending prompts and rendering repliesRepoDoc.py- Reference script for agent interaction
- Python 3.9 or later
azure-ai-projectsPython packageazure-identityPython package- Azure AI project endpoint and credentials
Install dependencies:
pip install -r requirements.txtOr install packages directly:
pip install azure-ai-projects>=2.1.0 azure-identity>=1.17.0server.py reads these environment variables if set:
AZURE_AI_PROJECT_ENDPOINT- Azure AI project endpoint URLREPO_DOC_AGENT_NAME- Agent name (default:RepoDoc)REPO_DOC_AGENT_VERSION- Agent version (default:7)REPO_DOC_HOST- Host address (default:127.0.0.1)REPO_DOC_PORT- Port number (default:8001)
Start the server:
python server.pyOpen the app in your browser:
http://127.0.0.1:8001
- The browser sends the user prompt to
POST /api/chat server.pyforwards the prompt to the Azure AI agent- The agent response returns to the server
- The browser renders the reply in the chat view
- Change the agent endpoint and name by editing environment variables or the defaults in
server.py - Update the UI in
static/index.html,static/style.css, andstatic/script.js - Add richer session or repo context support in
server.py
The app uses Azure Identity to authenticate with Azure AI. Ensure your local environment has valid Azure credentials available before running the server.