Get started building AI applications in minutes with the new AI App Template Gallery.

We’re thrilled to announce the launch of the AI App Template Gallery, a new resource designed to help you build and deploy AI applications. This gallery is a curated collection of ready-to-use application templates, creating a clear path to kickstart your AI projects with confidence and efficiency.

This collection includes fully working applications, complete with app code, AI features, infrastructure as code (IaC), configurable CI/CD pipelines with GitHub Actions, and an application architecture, ready to deploy to Azure. The collection also includes smaller templates that are building blocks hosting just a component you need, like keyless authentication or IaC.

How to deploy a template

Deploying a template from our AI App Template Gallery can be done in a few steps to get started:

1. Visit the gallery. Browse the available templates and select the one that best fits your project needs.

2. Open in GitHub Codespaces to deploy (or open the GitHub repo to view).

3. Follow the guidelines in the README to deploy the template with one command azd up into your development environment in as little as 5 minutes! Once you deploy, you can make changes to the app, then run azd deploy to push your changes. Remember, if you’re deploying a template to test, make sure to run azd down to clean up your resources.

Benefits for developers

The AI App Template Gallery offers numerous benefits for developers, including:

• Ease of use: The templates are built with the developer experience in mind, ensuring that even if you’re new to AI development you can deploy and customize them. The templates include:
  • Simple workflows: Deploy with one command azd up from the CLI and GitHub Actions set you up for iterative changes with azd deploy with your CI/CD pipeline.
  • Integration with GitHub Copilot: GitHub Copilot for Azure recommends and helps deploy AI App Templates for a seamless experience.
  • Comprehensive documentation: Each template comes with consistent README instructions, guiding developers through the deployment and customization process step-by-step.
• Accelerated time to market: Using prebuilt templates reduces the time required to set up and start projects, allowing developers to focus more on innovation and less on initial setup.
• Built-in security and reliability: The templates are built with a focus on quality, and repeatability.
  • Each template is configured to use Azure recommended practices including keyless authentication that reduces the risk of leaked API keys.
  • The templates are also crafted with recommended practices for programming language and LLM configurations.
  • The templates are regularly tested and used by thousands of developers across the globe.
• Flexibility and choice: The templates provide a choice of programming languages, models, frameworks, and services from Microsoft and leading AI toolchain ISVs. The templates can be used from within your preferred development environment including GitHub Codespaces, VS Code, and Visual Studio.

Templates build with leading AI ISVs

We have several templates that are built with your favorite technologies, including:

• Arize, a machine learning observability platform for ML practitioners to monitor, troubleshoot, and explain models. The Arize sample app demonstrates how you can monitor models and LLM applications, troubleshoot retrieval and tool execution issues, and improve the RAG pipeline.
• LangChain, a popular open-source framework that simplifies the creation of applications using LLMs. The serverless genAI assistant app (Serverless GenAI assistant with LangChain | AI app templates) shows how enterprise documents can be used to generate responses to user queries using the LangChain serverless application.
• LlamaIndex, a data framework for building LLM applications. There’s a template available in TypeScript and one in Python, with a full tutorial using the template. These two templates give you a GenAI chat application that uses your data, and sample data if you’re trying the template.
• Pinecone, a popular vector database that lets you create and scale vector indices for AI search and recommendation. The Pinecone Assistant sample app in Python and TypeScript helps you deploy your first Pinecone Assistant-based application. The template sets up an Azure Container App, linking to a Pinecone index and an OpenAI embedding model for storing and retrieving context for a RAG model.

Let us know what you think

We created the AI App Template Gallery to make building innovative AI solutions more efficient. Visit azure.github.io/ai-app-templates today to explore the gallery and find the perfect template for your next project!

We look forward to seeing the incredible applications you create using the AI App Template Gallery as a jumping off point. Let us know, on the azd discussion board, if there are templates you would like to see us build and add to the gallery.