This MCP server provides a comprehensive bridge to Cisco Webex Messaging for AI assistants. It exposes 52 tools categorized into Messages, Rooms, Teams, Memberships, People, Webhooks, and Enterprise features, enabling operations such as sending and editing messages, managing rooms and teams, handling memberships, working with webhooks, and accessing enterprise capabilities like room tabs and attachments. The server supports both STDIO and HTTP transport modes, so it can be used with local clients or HTTP-based MCP clients. To integrate with Claude Desktop or other MCP clients, you typically run the server in Docker (or locally in Node) and point your client to the appropriate MCP endpoint, then use the tool discovery features to learn available actions.

Core capabilities include:

• Messages: create_message, list_messages, edit_message, delete_message, get_message_details
• Rooms: create_room, list_rooms, get_room_details, update_room, delete_room
• Teams: create_team, list_teams, get_team_details, update_team, delete_team
• Memberships: create_membership, list_memberships, update_membership, delete_membership, create_team_membership, list_team_memberships
• People: get_my_own_details, list_people, get_person_details, create_person, update_person, delete_person
• Webhooks: create_webhook, list_webhooks, get_webhook_details, update_webhook, delete_webhook, list_events, get_event_details
• Enterprise: create_room_tab, list_room_tabs, get_room_tab_details, update_room_tab, delete_room_tab, create_attachment_action, get_attachment_action_details

To use the server in HTTP mode, start the server in HTTP mode (for example, npm run start:http) and then connect your MCP client to the provided endpoints (health and mcp). In STDIO mode, run the server as described by your deployment method and interact via the standard input/output transport.

Prerequisites:

• Docker installed on your machine (for containerized deployment)
• Node.js 18+ if you choose to run locally (optional)
• Webex API token from developer.webex.com

Installation steps (Docker-based):

1. Ensure you have Docker installed and running.

2. Create a Docker image for the MCP server if not already available: docker build -t webex-mcp-server .

3. Prepare environment variables in a .env file or provide them directly: WEBEX_PUBLIC_WORKSPACE_API_KEY=your_token_here WEBEX_USER_EMAIL=your.email@company.com WEBEX_API_BASE_URL=https://webexapis.com/v1

4. Run the MCP server in containerized mode (STDIO or HTTP as needed): // STDIO mode (default) docker run -i --rm --env-file .env webex-mcp-server

   // HTTP mode (exposed port 3001 by default) docker run -p 3001:3001 --env-file .env webex-mcp-server --http

5. If you prefer local Node execution, install dependencies and run directly: git clone <repository-url> cd webex-messaging-mcp-server npm install

   For STDIO mode

   node mcpServer.js

   For HTTP mode

   npm run start:http

Environment variables are required for authentication and API access. Ensure WEBEX_PUBLIC_WORKSPACE_API_KEY contains the token without the 'Bearer ' prefix. If you switch to STDIO mode, you can directly pipe input/output with your MCP client. The server supports CORS with mcp-session-id exposure for web clients and provides Server-Sent Events (SSE) for streaming responses in HTTP mode. If you encounter token expiry, renew tokens via the Webex developer portal and update WEBEX_PUBLIC_WORKSPACE_API_KEY in your environment. For Node.js environments with older Node versions, ensure fetch is available or polyfilled as described in the quick start notes.