ToolsCompaniesAbout

Microsoft Foundry

by Microsoft@microsoftRedmond, Washington, USANo ratings yet

Enterprise platform for building, grounding, and governing AI apps and agents

Visit azure.microsoft.com
Tool type
Unified AI app and agent platform
Made for
Developers, data scientists, IT admins, security teams
Runs on
Azure cloud portal and APIs; Foundry Local on-device
Foundry is accessed via the Foundry portal at ai.azure.com; Foundry Local runs models on-device and at the edge.
Last checked
28 Oct 2026
2 product sources

What it does

Microsoft Foundry takes in model choices, application code, business data sources, and tool connections, and produces deployed AI apps and agents with monitoring and governance attached. It covers the agent lifecycle: access to more than 11,000 models, Foundry Agent Service for hosting agents, grounding through Foundry IQ and Microsoft Graph, and pre-built or custom tools exposed over MCP. Foundry was formerly known as Azure AI Studio.

Teams typically start by picking a model, then wiring an agent to business data and actions, such as the 1,400+ prebuilt connections in Azure Logic Apps, including SAP, Salesforce, and Dynamics 365. The Foundry Control Plane sets the platform apart, integrating with Entra, Purview, and Defender and providing OpenTelemetry tracing, built-in evaluators for groundedness and safety, and runtime content protection. Developers can also work from Visual Studio Code with the Foundry Toolkit, and use Foundry Local to run models on-device and at the edge.

It suits developers, data scientists, IT administrators, and security teams at organizations building and scaling AI solutions, including startups and ISVs. You can explore the portal at ai.azure.com without an Azure account, but an Azure subscription is required before building agents. Billing is per service, so costs depend on the models, tools, and features you use.

What it can do

  • Access over 11,000 foundational, open, reasoning, multimodal, and industry-specific models
  • Fine-tune and distill models with minimal coding
  • Intelligent model routing that selects a model per task
  • Host agents with serverless containers and event-driven functions via Foundry Agent Service
  • Permission-aware grounding through Foundry IQ and Microsoft Graph
  • Run models and capabilities on-device and at the edge with Foundry Local

Works with

  • Visual Studio Code (Foundry Toolkit)
  • GitHub Copilot
  • Microsoft Copilot Studio
  • Microsoft 365 Copilot and Teams
  • Microsoft Fabric
  • Azure Cosmos DB
  • Azure Database for PostgreSQL
  • Azure Databricks
  • Azure Logic Apps
  • Microsoft Entra ID

Example workflows

Grounded customer-service agent
Illustrative example. Input: a customer question and an Azure-hosted product knowledge source. Steps: select a model in Foundry Models, ground the agent with Foundry IQ, add a business-system tool via Azure Logic Apps, and review traces in the Control Plane. Output: an agent that answers from the knowledge source and can call connected systems, with per-interaction tracing and safety evaluations.
Developer prototype in VS Code
Illustrative example. Input: an application codebase and a chosen model. Steps: install the Foundry Toolkit for Visual Studio Code, build and test the agent locally, then deploy it to Foundry Agent Service. Output: a deployed agent endpoint that the application calls, monitored by Foundry observability.

Before you choose

  • Pricing is per service and feature; the platform is free to explore, but consumed services are billed, so check the pricing page for current rates.
  • For Foundry Local pricing and availability, request a quote or check with Microsoft.
  • Using additional integrations such as Copilot Studio, Agent 365, or Fabric may add costs and governance setup beyond the base platform.

Listed under

Get started

No Azure account is needed to explore, but an Azure subscription is required to build agents.

  1. Visit ai.azure.com and explore the Foundry portal.
  2. Create or sign in to an Azure subscription before building agents.
  3. Choose a model from Foundry Models and connect any knowledge sources or tools needed for grounding.
Explore Foundry, then sign up for Azure

Plans and prices

Prices in USD

Platform is free to explore; you pay for consumed services. Azure offers a 30-day trial with credit. Quoted pricing is an estimate in US dollars and varies by agreement and currency. The Agent pre-purchase plan offers discounts for 1-year ACU commitments.

Usage-based (per service)Usage-basedNamed meters vary by service; see the Foundry pricing details page for current rates.
  • Foundry Models billed by model usage
  • Foundry Agent Service and Foundry Tools billed per service

30-day free trial.

Source: azure.microsoft.com/en-us/pricing/details/ai-foundry, checked 28 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 28 Oct 2026. Each source supports part of the overview; availability and prices can change.

Ratings and reviews

No ratings for Microsoft Foundry yet.

Be the first to rate it

What’s changed

Free account

See what’s changed at Microsoft Foundry

Launches, price moves and pivots, each with a date and a source. Free with an account.

Signals

Free account

See Microsoft Foundry’s signals

Launch traction, hiring and release activity, tracked over time. Free with an account.

Data and security

Free account

See how Microsoft Foundry handles your data

Training on your files, compliance, hosting and the models behind it. Free with an account.

Compare with similar tools

Free account

Compare Microsoft Foundry side by side

Prices, platforms and features against similar tools, with sources. Free with an account.

Where this comes from

How we research
Company website
azure.microsoft.com: product, pricing and policy pages
YC directory mirror
Batch and company name
Web search
Confirms the official site
Social accounts
Accounts the company links, and its posts on X
Human review
Checked by a person
knowledge_base

Last checked 28 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.