Skip to main content
Qwen Cloud home page
Docs
Best Practices

Integrate tools/MCP

Extend Token Plan with web search, code interpreter, and more

Token Plan lets you extend the capabilities of your AI programming tools with built-in tools and MCP services, which provide features like web search, code interpreter, and web scraping.

Tool overview

Token Plan provides two ways to integrate tools:
  • Built-in tools: The Responses API for the qwen3.6-plus model includes five built-in tools: web search, code interpreter, web scraping, search by image, and text-to-image search. When enabled, the model automatically calls the appropriate tool when needed.
  • MCP services: Other models, such as deepseek-v3.2, can gain tool capabilities through MCP services. This guide uses the web search MCP as an example, but the integration process is similar for other MCP services.

Billing

Built-in tools

Your Token Plan subscription covers the costs of the built-in tools for the qwen3.6-plus model. For specific pricing, refer to the model details page in the console.

MCP services

MCP services such as web search, code interpreter, and web scraping are available separately. Some services are offered free for a limited time and include a monthly free quota.

Usage

Use built-in tools for qwen3.6-plus

Set the model for your AI tool to qwen3.6-plus and ask questions directly in the chat. The model automatically calls the appropriate built-in tool based on your prompt:
ToolDescription
Web searchSearches the internet and generates answers based on the search results.
Code interpreterAllows the model to write and run Python code in a sandbox environment to solve complex problems, such as mathematical calculations and data analysis.
Web scrapingAccesses a specified URL and extracts its content to supply the large model with the information it needs.
Search by imageLets the model search the internet for visually similar images based on an input image, and then analyze and reason based on the search results.
Text-to-image searchLets the model search the internet for relevant images based on a text description, and then describe and reason based on the image content.

Integrate tools through MCP services

Other models can acquire tool capabilities through MCP services. The following sections use the web search MCP as an example to illustrate the integration process.

Prerequisites

You have obtained a Qwen Cloud API key. This is a general-purpose API key (in sk-xxx format), used to call MCP services. It is different from the dedicated API key for Token Plan (in sk-sp-xxx format).

Activate an MCP service

  1. Find the MCP service you need (for example, web search).
  2. Click Activate Now and confirm.
  3. After activating the service, get the following details:
    • Streamable HTTP Endpoint: The connection address for the MCP service.
    • API key: Your Qwen Cloud API key, used as the authentication key for the MCP service.

Integrate tools

Add the MCP service to your AI programming tool. The following sections use the web search MCP as an example. In the code samples, replace YOUR_API_KEY with your Qwen Cloud API key. To integrate other MCP services, replace the endpoint URL with the one for your service.
  • OpenClaw
  • OpenCode
  • Claude Code
  • Qwen Code
  • Kilo CLI
  • Kilo Code IDE
  1. Install MCPorter:
npm install -g mcporter
  1. Enable MCPorter:
openclaw config set skills.entries.mcporter.enabled true
  1. In ~/.openclaw/workspace, add the WebSearch MCP:
mcporter config add WebSearch https://dashscope-intl.aliyuncs.com/api/v1/mcps/WebSearch/mcp --transport http --header "Authorization=Bearer YOUR_API_KEY"
  1. Confirm the MCP is installed:
mcporter list
  1. Apply the configuration:
openclaw gateway restart
  1. Send Use mcporter to search for Qwen Cloud news to verify.