# What is Microsoft Foundry?

Microsoft's Azure platform for AI apps and agents, formerly Azure AI Foundry and Azure AI Studio: one resource to deploy models from many makers, build agents and govern both.

- Page: https://tecyfy.com/concepts/microsoft-foundry
- Subject: How AI works (https://tecyfy.com/concepts)
- Level: beginner
- Also called: Azure AI Foundry, Azure AI Studio, Foundry

Building with AI inside a large company means more than calling a model. Someone has to decide which models are allowed, who may use them, what data agents can reach, and how to see what they did. Microsoft Foundry is Azure's answer: one resource where all of that is set once.

## One resource, several jobs

A Foundry resource holds **projects**, and a project holds:

- **Model deployments.** More than 10,000 models are in the catalogue, from OpenAI, Anthropic, Meta, Microsoft and others. A deployment picks how the model is served: Global Standard (the default), Data Zone to keep processing in the US, EU or Asia Pacific, Provisioned for reserved capacity, or Batch at half price.
- **Agents.** *Prompt agents* are declared, not coded: instructions, a model and tools. *Voice* prompt agents answer spoken questions. *Hosted agents* are your own code, written with Microsoft Agent Framework, LangGraph or Semantic Kernel, packaged as a container and run by Foundry.
- **Tools and knowledge.** A toolbox adds built-in tools, retrieval and memory to agents.
- **Observability.** Traces and evaluations show what models and agents did.

Over all of it sits the governance: Entra ID sign-in, role-based access, content filters, private networking and Azure Policy, applied to every model and agent in the resource. That is the main reason to use it: a security team reviews one platform, not each project.

## A rename with real changes

The names have moved fast:

| Before | Now |
|---|---|
| Azure AI Studio, then Azure AI Foundry | Microsoft Foundry |
| Azure AI Services | Foundry Tools |
| Hub + Azure OpenAI + AI Services resources | One Foundry resource with projects |
| Assistants API | Responses API |
| A dated `api-version` on every call | The v1 API at `/openai/v1/` |

So "Azure OpenAI" is now one part of Foundry: OpenAI's models deployed and called through it. See [Azure OpenAI](https://tecyfy.com/concepts/azure-openai).

## How you reach it

Through the Foundry portal, the `azure-ai-projects` SDK (which hands you a standard OpenAI client), the Azure Developer CLI for hosted agents, VS Code, and coding agents through an [MCP](https://tecyfy.com/concepts/model-context-protocol) server. Calling a model is plain OpenAI SDK code pointed at your resource.

Our article [What is Microsoft Foundry](/blog/microsoft-foundry) lists every old name beside its new one and what to migrate first.

## Related

- [Azure OpenAI](https://tecyfy.com/concepts/azure-openai): OpenAI's models run by Microsoft inside Azure, now part of Microsoft Foundry. Same models and SDK as OpenAI's API, but your data stays with Microsoft, under your Azure agreement.
- [Agent](https://tecyfy.com/concepts/agent): A model that uses tools in a loop, deciding each next step from what the last one returned, until the task is done.
- [Large Language Model](https://tecyfy.com/concepts/large-language-model): A model trained on a huge amount of text to predict the next token. Further training turns it into an assistant that follows instructions.
- [Guardrail](https://tecyfy.com/concepts/guardrail): Checks placed around a model that stop or change what goes in and what comes out, such as a filter on user messages or an approval step before an agent acts.
- [Model Context Protocol](https://tecyfy.com/concepts/model-context-protocol): An open standard for connecting AI applications to tools and data. A system offered once as an MCP server works in any application that speaks MCP.

## Sources

- [What is Microsoft Foundry?](https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry)
- [Deployment types for Microsoft Foundry Models](https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/deployment-types)