Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
| Enabled for | Public preview | General availability |
|---|---|---|
| Users, automatically |
May 20, 2025 |
Dec 15, 2025 |
Business value
A remote Microsoft Dataverse model context protocol (MCP) server represents the next evolution in how AI systems and enterprise data interact within the Microsoft ecosystem. Unlike local MCP setups, which require developers or users to install SDKs, configure proxies, and manage environment access manually, a remote Dataverse MCP server runs as a centrally hosted, managed service. This shift streamlines integration, improves security, and enables scalable AI connectivity across teams, devices, and applications.
The remote Dataverse MCP server removes the friction of local setup and dependency management. All configuration, authentication, and environment mappings are managed centrally in the cloud. Organizations can onboard new users or environments instantly, reducing IT overhead and improving speed-to-adoption.
A simple sign-in or OAuth connection is enough to link all clients (like Claude, Copilot Studio, or Visual Studio Code extensions) to Dataverse. Ideal for distributed or hybrid teams, where users might access enterprise data from multiple devices or networks.
The remote Dataverse MCP server ensures uniform configuration and enterprise-grade security across Microsoft and other clients. Security policies and data boundaries can be centrally enforced through Power Platform admin center, ensuring compliance at scale.
Feature details
With this release, we are introducing advanced governance capabilities, new productivity tools, and updated billing policies for the Dataverse MCP server integrated with Microsoft Copilot Studio, Visual Studio Code, and Claude Desktop.
Key Updates:
- Governance model: We have introduced enhanced governance experience within Power Platform admin center to manage Dataverse MCP server. With the new governance framework, only Microsoft Copilot Studio will be enabled by default for existing Dataverse MCP server users.
- Configuration management: To enable additional clients like Visual Studio GitHub Copilot or Claude, configure them through Power Platform admin center settings. To disable the Dataverse MCP server, go to Power Platform admin center environment settings and turn off the feature.
- Connecting to Visual Studio GitHub Copilot: We have introduced a simplified experience to interact with Dataverse MCP server in Visual Studio Code through remote MCP configuration.
- Billing and licensing: Effective December 15, 2025 usage of the Dataverse MCP Server will be subject to billing.
- New capabilities: Access the newly introduced tools and enhanced features designed to optimize development and operational workflows.
- search_data - Search structured and unstructured data.
- search - Search table schemas and business skills by keyword.
- create_record - Inserts a new row into a Dataverse table and returns the GUID.
- update_record - Updates an existing row in a Dataverse table.
- delete_record - Delete a row, only after explicit user approval.
- create_table - Creates a new table with a specified schema.
- update_table - Modifies schema or metadata of an existing table.
- delete_table - Deletes a table from Dataverse, only after explicit user approval.
- read_query - Run supported Dataverse SQL SELECT queries.
- describe - Get details from search results for tables, records, schemas, skills, and apps.
- upsert_skill - Add or update a Dataverse skill or playbook.
- delete_skill - Delete a Dataverse skill or playbook by name.
- init_file_upload - Generate a SAS URL for file upload.
- commit_file_upload - Commit a file upload.
- file_download - Generate a SAS URL for file download.
Geographic areas
Visit the Explore Feature Geography report for Microsoft Azure areas where this feature is planned or available.
Language availability
Visit the Explore Feature Language report for information on this feature's availability.
May 20, 2025