# Tencent TRTC MCP for AI Agents AI Agent Connect > Tencent TRTC lets you give your AI agent the keys to your live-streaming infrastructure. It handles the heavy lifting of managing video rooms, starting cloud recordings, and monitoring call quality across your global platform. Instead of manually juggling dashboards, your agent takes over the session management. ## Overview - **Category:** industry-titans - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_GThlN2sMLsphbcxAIZc1i54N6RbpaNYdpaefU2AR/ai-agent-connect - **Tags:** real-time-communication, live-streaming, video-calls, session-management, cloud-recording, telemetry ## Description Imagine you're running a massive live-streaming platform where hundreds of rooms are active at once. Usually, this means a team of moderators and engineers constantly checking for bad actors, starting recordings, and troubleshooting lag. This Connector changes that by letting your AI agent handle the grunt work directly. You can give your agent the power to spot and kick out troublemakers, manage complex cloud recording tasks, and even toggle on transcoding for better stream mixing. It's about moving from reactive troubleshooting to proactive management. When you connect this through the Vinkius catalog, your AI becomes a first-class citizen in your infrastructure stack. It can parse through call metrics to find out why a specific user's video is stuttering or ensure that every high-profile event gets recorded to the cloud without a human having to click start every single time. It turns your real-time communication setup into a programmable environment where your agent can actually do work rather than just talk about it. ## Tools ### describe_call_detail_info Pulls granular metrics on call quality like latency and dropped frames. This lets your agent see exactly why a stream is lagging. ### start_mcu_mix Turns on MCU mix transcoding for a room to improve stream quality. This helps viewers get a smoother experience. ### stop_cloud_recording Ends an active cloud recording task to save space and costs. Use this to stop recording as soon as an event finishes. ### stop_mcu_mix Turns off MCU mix transcoding when it is no longer needed for a session. This helps you manage your resource usage. ### describe_user_info Gets a list of users in a specific call session using a unique identifier. This helps your agent know exactly who is in the room. ### describe_room_info Provides the current status and details of a specific TRTC room. Use this to get a snapshot of the room's health. ### describe_trtc_usage Pulls aggregated statistics on your overall TRTC usage. This is great for keeping an eye on your platform's health and costs. ### dismiss_room Ends a live TRTC room session immediately to clear the space. Use this to shut down old sessions quickly. ### remove_user Kicks a specific participant out of a live room. This is the primary way to handle bad actors in real time. ### remove_user_by_str_room_id Removes users from a room using a string-based room identifier. It provides a flexible way to clear out participants. ### start_cloud_recording Begins a high-definition cloud recording for a specific room. This ensures your content gets saved to your storage. ## Prompt Examples **Prompt:** ``` Kick the user who is spamming the chat from room 502. ``` **Response:** ``` I've identified the user and used remove_user to disconnect them from room 502. The session is now clear of the spammer. ``` **Prompt:** ``` Is the recording for the main stage active right now? ``` **Response:** ``` I checked the status for the main stage room. The cloud recording is currently active. I can also start or stop it for you if needed. ``` **Prompt:** ``` Why is the video lagging for the users in the VIP room? ``` **Response:** ``` I've pulled the call details for the VIP room. It looks like there's a high rate of dropped frames (approx 15%) for some participants, which suggests a network stability issue rather than a server-side problem. ``` ## Capabilities ### Kick bad actors Automatically remove problematic users from live sessions to keep your community safe. ### Manage cloud recordings Start and stop high-definition recordings to your cloud storage on demand. ### Handle stream mixing Toggle MCU mix transcoding to ensure smooth video playback for all participants. ### Monitor call quality Pull real-time data on dropped frames and network performance to fix lag. ### Track active users Get a live count of who is in a room to manage capacity and engagement. ### Retrieve session details Pull specific room info to help your agent understand the current environment. ## Use Cases ### Trolling prevention An AI agent monitors a live room and uses remove_user when it detects a user violating community guidelines. ### Automated archives A user tells their agent to "Record the next hour of this stream," and the agent calls start_cloud_recording. ### Lag diagnosis A viewer complains about stuttering, so the agent calls describe_call_detail_info to see if it's a dropped frame issue. ### Event cleanup After a webinar ends, the agent uses dismiss_room to shut down the session and free up resources. ## Benefits - Automated moderation: Use remove_user to instantly clear out trolls without a human moderator needing to be present 24/7. - Hands-off recording: Use start_cloud_recording and stop_cloud_recording to ensure every important moment is saved to your cloud storage. - Better stream quality: Toggle start_mcu_mix and stop_mcu_mix to give viewers the best possible video experience during peak loads. - Real-time troubleshooting: Use describe_call_detail_info to let your agent identify exactly where network lag is happening. - Scalable room management: Use dismiss_room to clear out old sessions and manage high-volume traffic across your entire platform. - Data-driven insights: Use describe_trtc_usage to keep a pulse on your infrastructure costs and overall platform health. ## How It Works The bottom line is you get an AI agent that can actually manage your live-streaming infrastructure instead of just watching it. 1. Subscribe to the Tencent TRTC MCP and grab your SecretId, SecretKey, and SdkAppId from your developer console. 2. Add the Connector to your AI client and provide the credentials so it can authenticate with your Tencent account. 3. Give your agent instructions to manage your rooms, and it will start executing commands like kicking users or starting recordings. ## Frequently Asked Questions **Can the Tencent TRTC MCP help me moderate my live streams?** Yes, it allows your AI agent to automatically remove users from a room. This means you can keep your community safe without needing a human to watch every single interaction 24/7. **How does this help with cloud recordings?** You can tell your agent to start or stop recordings for specific rooms. It handles the interaction with Tencent's cloud storage so you don't have to manually trigger it. **Can my AI agent check why my video stream is lagging?** Yes, it can pull granular metrics like dropped frames and latency. This helps you diagnose whether the issue is on the user's end or a broader network problem. **Does this work for managing massive amounts of rooms?** It's designed for that. You can use it to dismiss rooms, manage MCU mixing for better quality, and get usage stats to keep your infrastructure running smoothly. **Is this the right tool for a video conferencing app?** If your app uses Tencent TRTC for its real-time communication, then yes. It gives your AI agent the ability to manage the actual sessions and room environments. **What happens to my recordings when I use this?** The agent interacts with your existing Tencent VOD setup. It just handles the start and stop commands for you automatically. **Does this support Tencent Cloud API v3 signature?** Yes! The Connector abstraction layer perfectly encapsulates and executes the TC3-HMAC-SHA256 signature algorithm natively so you don't have to code it. Just provide `SecretId` and `SecretKey` safely.