Integrate AI with Microsoft Teams Communications

Microsoft Teams MCP Server is an MCP server for Microsoft Teams integration, giving an AI assistant the ability to read messages, start new threads, reply to existing threads, and mention members directly in a Teams channel.

Use it when you want an AI assistant to post updates to a Teams channel, reply to an existing conversation thread, read channel messages or thread replies, or list a channel's team members, all with the ability to at-mention specific people. It requires a Microsoft Teams account set up with an MS Entra ID application, TEAMS_APP_ID, TEAMS_APP_PASSWORD, and tenant configuration, plus a specific Team ID and Channel ID, so it depends on Azure/Entra app registration work being done first; the project's own setup document is explicitly not a step-by-step guide but a pointer to what you'll need. It is scoped to a single configured Team and Channel per running instance, not a general cross-tenant Teams browsing tool.

Thread and message tools cover starting a new thread in a channel with a title and content, mentioning users as needed; updating an existing thread with reply messages, again with mentions; and reading thread replies. Channel-level tools list the members of a team and read channel messages directly. Authentication runs through an MS Entra ID app registration, configured via TEAMS_APP_ID, TEAMS_APP_PASSWORD, TEAMS_APP_TYPE, SingleTenant or MultiTenant, and TEAMS_APP_TENANT_ID for single-tenant setups, alongside TEAM_ID and TEAMS_CHANNEL_ID identifying the specific Team and Channel to operate on.

Requires uv and Python 3.10, plus a properly configured Microsoft Teams account and Azure app registration. Clone and set up dependencies:

git clone [repository-url]
cd mcp-teams-server
uv venv
uv sync --frozen --all-extras --dev

Set the required environment variables, or use the provided sample.env as a template, then start it with uv run mcp-teams-server. A pre-built Docker image is also available from ghcr.io, docker pull ghcr.io/inditextech/mcp-teams-server:latest, or build your own with docker build . -t inditextech/mcp-teams-server and run it with docker run --env-file .env -it inditextech/mcp-teams-server.

Integration testing has its own environment variables, TEST_THREAD_ID, TEST_MESSAGE_ID, and TEST_USER_NAME, and can be run with uv run pytest -m integration.

Teams within organizations using Microsoft Teams who want an AI assistant to post updates, reply to threads, and read channel activity directly in a specific Teams channel, instead of switching to the Teams app for every message. It is released under the Apache-2.0 License.