This repository contains sample projects, primarily written in Python, that demonstrate how to build and deploy AI applications on Azure App Service. Each sample is designed to showcase best practices for integrating AI capabilities and deploying scalable web apps to Azure.
# 1) Clone the repo
git clone https://github.com/Azure-Samples/appservice-ai-samples.git
cd appservice-ai-samples
# 2) Pick a sample
cd <sample-folder> # e.g., lowlight-enhancer
# 3) Initialize and deploy
azd init
azd upPrereqs: An Azure subscription and the Azure Developer CLI (
azd) installed.
-
langchain-fastapi-chat: A FastAPI web app using LangChain and Azure OpenAI (gpt-4o) for conversational AI, featuring streaming responses and summaries. Includes Bicep templates for easy deployment to App Service.
-
image-caption-app: A Streamlit app that generates natural-language captions for uploaded images using an LLM + vision pipeline. Demonstrates a simple, user-friendly UI with drag-and-drop, promptable captioning, and an App Service–ready structure.
-
gpt-oss-20b-sample: A lightweight chat app that calls the GPT-OSS-20B model with a minimal web UI and clear, configurable prompts. Shows a clean scaffolding for inference calls, streaming responses, and environment-based configuration
-
lowlight-enhancer: A tiny Flask + OpenCV web app that improves dim photos (CLAHE → gamma → brightness → subtle saturation). CPU-friendly, no heavyweight model required; instant before/after preview in the browser and a minimal code path that’s easy to extend.
Useful documentation and related resources: