ZEGO / 即构科技 MCP. Manage streams and audit users from a single chat window.

You're connecting your AI agent directly to ZEGO’s global RTC services. This server gives you full command over real-time comms infrastructure, letting your agent handle complex operations—like monitoring dozens of live rooms or auditing bandwidth usage—without needing a human to touch any dashboard. Your agent executes these actions through simple conversational commands, giving it direct control over video/audio streams and user presence.

Room Management & Presence Tracking

The server lets your agent find all active communication hubs using list_rooms, which returns every room currently running on the system. To audit a specific session, you can call get_room_users to pull a detailed list of every user ID and metadata for people logged into that particular room. You'll also use get_online_count to get a single number: the current total count of all users connected across the platform.

When your agent needs to check if specific folks are online or offline, it runs through check_user_status, which checks the live connection status for multiple specified accounts.

Media Stream and Usage Control

Your agent can list every single active audio and video media stream running inside a room using get_room_streams. If you need to stop a bad feed or terminate an unwanted broadcast, it runs stop_media_stream, forcing the termination of that specific media stream. For financial auditing, your agent pulls historical data via get_usage_stats.

This tool gathers service usage statistics, giving you total minutes consumed by audio or video over defined time periods.

Moderation and Enforcement

If a user breaks rules, your agent can kick them instantly using kick_room_user; this executes a command that immediately removes the specified user from their current room. To keep tabs on participation, you'll use get_room_users to retrieve the full list of participants in any given session.

Operational Workflow Summary

When your agent runs through these tools, it builds an operational overview for you. You can start by finding all active rooms with list_rooms. Then, if a user needs to see who's there, the agent uses get_room_users or get_online_count. If performance is bad, it checks the streams with get_room_streams and may force-stop them using stop_media_stream.

For accountability, you can check usage minutes with get_usage_stats, verify user presence using check_user_status, or enforce rules by removing users with kick_room_user. This whole setup lets your AI client manage the entire RTC lifecycle from one API source. You don't have to jump between technical dashboards; you just talk to your agent, and it handles the heavy lifting on your communication backend.