Integrate AI with Microsoft Teams Communications

MCP Teams Server is a Model Context Protocol server for Microsoft Teams that gives AI assistants the ability to read channel messages, start new threads, reply to existing threads, list channel team members, and mention users - covering the core read/write actions of a Teams channel.

Use it when you want an AI assistant to monitor or participate in a specific Microsoft Teams channel: summarizing conversations, posting updates, replying to threads, or mentioning teammates. It requires a Microsoft Teams account with an MS Entra ID application registered in Azure (app ID, client secret, tenant ID) plus a specific team and channel ID, so it is scoped to one configured channel rather than acting across an entire Teams tenant.

The server exposes tools to start a thread in a channel with a title and contents (mentioning users), update existing threads with replies (also with mentions), read thread replies, list channel team members, and read channel messages. It defaults to stdio transport, supports streamable-http for HTTP deployments via a --transport flag, and keeps the legacy sse transport available for older clients. The project's own integration test suite exercises these tools against a live Teams tenant, configured via TEST_THREAD_ID, TEST_MESSAGE_ID, and TEST_USER_NAME environment variables and run with uv run pytest -m integration.

Clone the repository and install with uv (requires the uv package manager and Python 3.12):

git clone https://github.com/InditexTech/mcp-teams-server
cd mcp-teams-server
uv venv
uv sync --frozen --all-extras --dev

Configure TEAMS_APP_ID, TEAMS_APP_PASSWORD, TEAMS_APP_TENANT_ID, TEAM_ID, and TEAMS_CHANNEL_ID as environment variables (a sample .env template is provided), then start it with uv run mcp-teams-server. A pre-built Docker image is also published at ghcr.io/inditextech/mcp-teams-server:latest, pullable directly or buildable from source with docker build . -t inditextech/mcp-teams-server; run it with docker run --env-file .env -it inditextech/mcp-teams-server, adding -p 8000:8000 and --transport streamable-http for HTTP deployments. A dedicated guide documents how to connect an LLM client to the running server. The project is Apache-2.0 licensed, maintained by INDITEX.

Teams and organizations using Microsoft Teams for internal communication who want an AI assistant to keep a specific channel updated, triage incoming messages, or participate in threaded discussions without a human manually relaying every message. The project also publishes SonarCloud quality metrics (bugs, maintainability, reliability) and an OpenSSF Scorecard rating for teams vetting it before deployment.