Overview
This document will guide you on how to configure Tencent Cloud SDK MCP Server in AI editors such as Cursor, TRAE, and VSCode. Once configured, the AI will possess full-chain live stream development capability. You just need to input natural language instructions to auto-build core features like basic push/pull stream (go live/view/list), real-time interaction (co-streaming/PK), and entertainment effects (bullet screen/gift/beauty effect/audio effect), saying goodbye to complex coding.
Core Capability Panoramic View
Basic push/pull stream: anchor starts live streaming, audience viewing, live list.
Real-time interaction: multi-person microphone connection, audience list, anchor connection, cross-room PK.
Entertainment special effects: beauty filter, gift rewards, real-time bullet screen, fun sound effects.
Step 1: Environment Preparation
1. Development Environment Preparation
Ensure your local development environment has the recommended version of Node.js and npm (Node Package Manager) installed.
Download and install:
Recommended version: Recommend using Node.js LTS (long-term support) version, typically in the 18.x or 20.x series.
Verify installation:
After installation, open the command window (also called command prompt, terminal) on your computer and run the following command to verify the version:
Windows systems: Press the Win + R keys, input cmd, then press Enter to open.
macOS: Press Command + space bar, input Terminal or Terminal, then press Enter to open.
If the version number is displayed correctly, the environment is ready.
If the prompt shows no related commands or other errors, provide the complete error information screenshot or text content to the AI so that the solution can be found more efficiently.
2. Get Application Information
Refer to Activate Service to claim the TUILiveKit trial version, and obtain the SDKAppID and secretKey on the Application Management page. These will be used as environment variables when configuring MCP. Step 2: MCP Tool Configuration
Before using the MCP tool, you need to configure it in the code editor (such as Cursor, TRAE, CodeBuddy, VSCode) configuration panel. It is recommended to set it to the root directory of the current project.
Cursor Configuration Guide
CodeBuddy Configuration Guide
VSCode Configuration Guide
2. Create a .cursor/mcp.json file in the project root directory, update the mcpServers configuration to mcp.json, and save it.
{
"mcpServers": {
"tencentcloud-sdk-mcp": {
"command": "npx",
"args": ["-y", "@tencentcloud/sdk-mcp"],
"env": {
"SDKAPPID": "YOUR_SDKAPPID",
"SECRETKEY": "YOUR_SECRET_KEY"
}
}
}
}
Note:
Please save promptly your modified configuration file:
For Windows users: Ctrl + S, for Mac users: Command + S.
3. When saving the file, the figure below will pop up. Select Enable to enable the MCP tool.
4. Configuration complete. Check whether the MCP tool is successfully loaded in the Settings - Tools & MCP option in Cursor.
2. Click Settings > MCP > +Add > Manually Add.
3. Click Original Configuration (JSON) to open the configuration file, update the mcpServers configuration in the configuration file, and save it.
{
"mcpServers": {
"tencentcloud-sdk-mcp": {
"command": "npx",
"args": ["-y", "@tencentcloud/sdk-mcp"],
"env": {
"SDKAPPID": "YOUR_SDKAPPID",
"SECRETKEY": "YOUR_SECRET_KEY"
}
}
}
}
Note:
Please save promptly your modified configuration file:
For Windows users: Ctrl + S, for Mac users: Command + S.
4. Select the editor's default Builder with MCP or add MCP to your custom Builder.
2. Click Settings > Add MCP to open settings.json and configure MCP.
3. Update the mcpServers configuration in the settings.json file and save it.
{
"mcpServers": {
"tencentcloud-sdk-mcp": {
"command": "npx",
"args": ["-y", "@tencentcloud/sdk-mcp"],
"env": {
"SDKAPPID": "YOUR_SDKAPPID",
"SECRETKEY": "YOUR_SECRET_KEY"
}
}
}
}
Note:
Please save promptly your modified configuration file:
For Windows users: Ctrl + S, for Mac users: Command + S.
4. Check whether the MCP tool successfully loads tools.
1. Install the VS Code CodeBuddy plug-in.
2. Once installed successfully, find the CODEBUDDY entry in the sidebar, click the MCP market icon > installed > configure MCP Server.
3. Update the mcpServers configuration in the configuration file and save it.
{
"mcpServers": {
"tencentcloud-sdk-mcp": {
"command": "npx",
"args": ["-y", "@tencentcloud/sdk-mcp"],
"env": {
"SDKAPPID": "YOUR_SDKAPPID",
"SECRETKEY": "YOUR_SECRET_KEY"
}
}
}
}
Note:
Please save promptly your modified configuration file:
For Windows users: Ctrl + S, for Mac users: Command + S.
4. Check whether the MCP tool successfully loads tools after saving succeeded.
Step 3: AI-Assisted Integration (Prompt Guide)
Preparing the Project Environment
New project: Select to create the corresponding platform's new project via Android Studio/Xcode/Flutter CLI based on your development platform, and use the AI editor to open it.
Existing project: Open your project directory directly in the AI editor.
Once configured, you just need to enter natural language instructions in the Chat panel. The AI will automatically identify the intent and call the tool to generate code.
Prompt content skill:
To obtain the best results, recommend using the structure of [action] + [core SDK] + [platform] + [feature] + [UI/detail requirements].
Note: When executing a long task, the AI may pause midway to ask for confirmation. Please click "Continue" or manually confirm to continue execution.
Seamless Integration of Basic Live Streaming Feature
Use cases: Project initialization, rapid deployment of push/pull streams and live list.
Recommended prompts:
1. Please use AtomicXCore to implement live stream publishing on the Android platform. // You can add attachment conditions behind, such as expected UI style, code location, code style, resource path, etc.
2. Please use AtomicXCore to implement audience viewing on the Android platform.
3. Please use AtomicXCore to implement live list on the Android platform.
4. Please use AtomicXCore to implement live stream publishing, live list, and audience viewing on the Android platform.
1. Please use AtomicXCore to implement live stream publishing on the iOS platform.
2. Please use AtomicXCore to implement audience viewing on the iOS platform.
3. Please use AtomicXCore to implement live list on the iOS platform.
4. Please use AtomicXCore to implement live stream publishing, live list, and audience viewing on the iOS platform.
1. Please use AtomicXCore to implement live stream publishing on the Flutter platform.
2. Please use AtomicXCore to implement audience viewing on the Flutter platform.
3. Please use AtomicXCore to implement live list on the Flutter platform.
4. Please use AtomicXCore to implement anchor starts live streaming, live list, and audience viewing on the Flutter platform.
Integrating Interactive Capability
Use cases: audience list, audience mic connection, anchor connection, cross-room PK, and other features.
Recommended prompts:
1. Please use AtomicXCore to implement audience list on the Android platform.
2. Please use AtomicXCore to implement audience mic connection on the Android platform.
3. Please use AtomicXCore to implement anchor connection on the Android platform.
4. Please use AtomicXCore to implement anchor PK Function on the Android platform.
1. Please use AtomicXCore to implement audience list on the iOS platform.
2. Please use AtomicXCore to implement audience mic connection on the iOS platform.
3. Please use AtomicXCore to implement anchor connection on the iOS platform.
4. Please use AtomicXCore to implement anchor PK Function on the iOS platform.
1. Please use AtomicXCore to implement audience list on the Flutter platform.
2. Please use AtomicXCore to implement audience mic connection on the Flutter platform.
3. Please use AtomicXCore to implement anchor connection on the Flutter platform.
4. Please use AtomicXCore to implement anchor PK Function on the Flutter platform.
Integrating Entertainment with Elastic Scalability
Use cases: beauty filter, gift rewards, real-time bullet screen, fun sound effects and other features.
Recommended prompts:
1. Please use AtomicXCore to implement the beauty effect feature on the Android platform.
2. Please use AtomicXCore to implement the sound effect feature on the Android platform.
3. Please use AtomicXCore to implement the danmaku function on the Android platform.
4. Please use AtomicXCore to implement the gift function on the Android platform.
1. Please use AtomicXCore to implement the beauty effect feature on the iOS platform.
2. Please use AtomicXCore to implement the sound effect feature on the iOS platform.
3. Please use AtomicXCore to implement the danmaku function on the iOS platform.
4. Please use AtomicXCore to implement the gift function on the iOS platform.
1. Please use AtomicXCore to implement the beauty effect feature on the Flutter platform.
2. Please use AtomicXCore to implement the sound effect feature on the Flutter platform.
3. Please use AtomicXCore to implement the danmaku function on the Flutter platform.
4. Please use AtomicXCore to implement the gift function on the Flutter platform.
Step 4: Running and Verification
Run the project: Click the Run button in the IDE.
Error fix: If there is a compilation error, copy the error information directly to the AI. It will analyze the error log and automatically fix the code.
FAQs
Whether There Is a Recommended Large Model?
Claude 3.5 Sonnet is highly recommended. Compared with other models, it delivers the best performance in understanding code logic, handling complex context, and accuracy in MCP tool calls.
How to Ensure AI Can Accurately Match to MCP Tool?
If AI is found not calling tools or manually fabricating code, it can be forcibly triggered in the following ways:
Explicitly declare (quickest): At the beginning of the Prompt, add "Please use MCP tool" to force AI to enter tool invocation method.
Example: "Please use MCP tool and implement the beauty effect feature on the Android platform with AtomicXCore."
What to Do If AI-Generated UI/Code Does Not Meet Expectations
Recommendation adoption of "incremental constraint method" for fine-tuning. Try using [basic instruction] + [UI/detail requirements] + [code specification] structure.
How to Manually Install MCP Server
If you need to manually install MCP Server, please execute the following commands in the terminal:
npx -y @tencentcloud/sdk-mcp
How to Resolve Folder Access Permission Issues during AI Project Creation
If you encounter a local directory access permission issue, recommend using sudo privileges to assign write permission to the directory.
Login Failure for UserSig Generated by MCP Tool, What to Do
Login failure with UserSig generated by the MCP tool may be because characters were modified during auto-write in certain IDEs. You can let the AI regenerate or manually copy the UserSig to the code.