Cursor in Microsoft Teams

Cursor Comes to Microsoft Teams

Cursor extended its cloud agent platform to Microsoft Teams on May 11, 2026, enabling development teams to delegate coding tasks without ever leaving their chat environment. By mentioning @Cursor in any Teams channel, team members can assign work to a cloud agent or surface information from Cursor directly inside Teams.

How It Works

The integration operates through Cursor's existing cloud agent infrastructure. When a team member mentions @Cursor in a channel, the agent reads the entire conversation thread to build context before taking any action. It then automatically selects the appropriate repository and model based on the contents of the prompt and prior agent activity — removing the need to manually configure which codebase or model to use.

Once the agent has gathered enough context, it implements a solution and opens a pull request for the team to review. The entire workflow — from task delegation to PR creation — happens asynchronously inside Teams, so engineers do not need to switch to the IDE to kick off a task.

Getting Started

The integration is available through the Cursor dashboard's integrations section. Teams and Enterprise plan users can enable it directly from account settings, with official documentation available in the Cursor docs under Microsoft Teams integration.

Why It Matters

This release continues Cursor's strategy of embedding cloud agents into the places where development teams already communicate. Having previously integrated with Slack for async task delegation, the Teams integration brings the same capability to the large portion of enterprise development organizations that standardize on Microsoft's collaboration suite. The ability to delegate, review, and discuss code without switching tools reduces friction in the development loop and makes Cursor's agents accessible to anyone on the team — not just developers with the IDE open.

Bug Fixes and Improvements

• Edge cases in cloud agent timing and hydration have been resolved to improve session reliability.
• Improved model selection logic based on recent agent activity history.