Microsoft Foundry
BetaManage Microsoft Foundry AI resources — agents, evaluations, model catalog/deployments, and connections — via Entra ID.
- Microsoft
- Azure
- AI Platform
- Authentication
- OAuth 2.0
- Hosting
- Vendor-hosted endpoint
- Required scopes
- 1
- Data actions
- Read and write
How it works
Every agent request passes through Ferentin policy before it reaches Microsoft Foundry.
Ferentin's Microsoft Foundry MCP server connects agents through Microsoft Entra ID (OAuth On-Behalf-Of), with 38 tools spanning agent lifecycle, evaluations and datasets, the model catalog and deployments, project connections, and prompt optimization. Operations run with the signed-in user's Azure RBAC permissions.
- AI agentAsks for something
- Ferentin policies
- Entra ID OAuth
- On behalf of
- Azure rbac
- No API keys
- Microsoft Foundry MCP serverAuthorized connection
- Governed resultMasked, logged, returned
What agents can do
38 tools · you choose which ones agents may callSearch and browse what exists, without opening the content itself.
- model_catalog_list
- model_quota_list
- project_connection_list
- project_connection_list_metadata
Permissions requested
The OAuth 2.0 scopes Microsoft Foundry asks you to approve when you connect.
| Scope | Access |
|---|---|
| https://ai.azure.com/.default | Not rated |
Before you connect
- Public preview: no SLA; the tool set may change without notice. Not recommended for production.
- Microsoft Entra does not support Dynamic Client Registration (DCR) for this server — each tenant registers its own Entra app.
- Auth is OAuth On-Behalf-Of: tools execute with the signed-in user's permissions (Reader = read tools, Contributor = writes). No API keys.
- The endpoint is public-only — Foundry resources behind Azure Private Link are not reachable through this server.
1 more technical note
- The OAuth scope (https://ai.azure.com/.default) and the 'organizations' authority are best-effort from current docs; verify against your Entra app registration.
About Microsoft Foundry
Microsoft Foundry (formerly Azure AI Foundry) is Microsoft's platform for building, evaluating, and deploying AI agents and models.
Frequently asked about Microsoft Foundry
- What can AI agents do with Microsoft Foundry through Ferentin?
- Ferentin exposes 38 Microsoft Foundry MCP tools to your AI agents, grouped as Discover (4 tools), Access (17 tools) and Manage (15 tools). Every call is checked against your Ferentin policies before it reaches Microsoft Foundry.
- How is Microsoft Foundry access secured?
- Requests through the Microsoft Foundry integration are covered by entra ID OAuth, on behalf of, azure rbac and no API keys. These controls are applied by Ferentin, not by Microsoft Foundry, so they hold across every connected agent.
- What permissions does the Microsoft Foundry integration require?
- The Microsoft Foundry integration requests 1 OAuth 2.0 scope: https://ai.azure.com/.default
- How do I connect Microsoft Foundry to Ferentin?
- Open the Ferentin admin console, add Microsoft Foundry as a new connection and authenticate with OAuth 2.0. You then choose which permissions and which tools your agents may use, and the connection goes live for everyone covered by your policies.
- Is the Microsoft Foundry integration generally available?
- The Microsoft Foundry integration is currently beta on Ferentin.
Related integrations
3
Microsoft Learn
Semantic search over official Microsoft/Azure docs and code samples — public and read-only.
Google Vertex AI
Google Cloud's enterprise AI platform with managed Gemini models, fine-tuning, and MLOps capabilities.
AWS Knowledge
Ground agents in AWS documentation, API references, best practices, and regional availability — public and read-only.