Fetch Live Web Data for AI Agents via Web Extractor - Model Studio

How it works

Include web_extractor (and typically web_search) in the tools array of your API request, along with a prompt that references a URL or topic.
The model identifies URLs, fetches content, and appends it to the context as input tokens.
The model generates a response grounded in the retrieved content.

Extracted content increases input tokens. See Billing for pricing details.

Prerequisites

You need:

An API key exported as DASHSCOPE_API_KEY (see Configure API key through environment variables)

Quick start

Use the Responses API to call web extractor. The examples use web_search, web_extractor, and code_interpreter with qwen3-max-2026-01-23 in thinking mode.

First, get an API key and export the API key as an environment variable.

Python

import os
from openai import OpenAI

client = OpenAI(
    # If the environment variable is not configured, replace with: api_key="sk-xxx"
    api_key=os.getenv("DASHSCOPE_API_KEY"),
    base_url="https://dashscope-intl.aliyuncs.com/api/v2/apps/protocols/compatible-mode/v1"
)

response = client.responses.create(
    model="qwen3-max-2026-01-23",
    input="Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it",
    tools=[
        {"type": "web_search"},
        {"type": "web_extractor"},
        {"type": "code_interpreter"}
    ],
    extra_body={
        "enable_thinking": True
    }
)

# Uncomment to view intermediate output
# print(response.output)
print("=" * 20 + "Response" + "=" * 20)
print(response.output_text)

# Print tool invocation count
usage = response.usage
print("=" * 20 + "Tool Invocation Count" + "=" * 20)
if hasattr(usage, 'x_tools') and usage.x_tools:
    print(f"Web Extractor invocations: {usage.x_tools.get('web_extractor', {}).get('count', 0)}")

Node.js

import OpenAI from "openai";
import process from 'process';

const openai = new OpenAI({
    // If the environment variable is not configured, replace with: apiKey: "sk-xxx"
    apiKey: process.env.DASHSCOPE_API_KEY,
    baseURL: "https://dashscope-intl.aliyuncs.com/api/v2/apps/protocols/compatible-mode/v1"
});

async function main() {
    const response = await openai.responses.create({
        model: "qwen3-max-2026-01-23",
        input: "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it",
        tools: [
            { type: "web_search" },
            { type: "web_extractor" },
            { type: "code_interpreter" }
        ],
        enable_thinking: true
    });

    console.log("====================Response====================");
    console.log(response.output_text);

    // Print tool invocation count
    console.log("====================Tool Invocation Count====================");
    if (response.usage && response.usage.x_tools) {
        console.log(`Web Extractor invocations: ${response.usage.x_tools.web_extractor?.count || 0}`);
        console.log(`Web Search invocations: ${response.usage.x_tools.web_search?.count || 0}`);
    }
    // Uncomment to view intermediate output
    // console.log(JSON.stringify(response.output[0], null, 2));
}

main();

curl

curl -X POST https://dashscope-intl.aliyuncs.com/api/v2/apps/protocols/compatible-mode/v1/responses \
-H "Authorization: Bearer $DASHSCOPE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
    "model": "qwen3-max-2026-01-23",
    "input": "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it",
    "tools": [
        {"type": "web_search"},
        {"type": "web_extractor"},
        {"type": "code_interpreter"}
    ],
    "enable_thinking": true
}'

Running the code above produces a response similar to:

====================Response====================
Based on the official Alibaba Cloud Model Studio documentation, here is a summary of the **Code Interpreter** feature:

## Overview

...

> **Source**: Alibaba Cloud Model Studio Official Documentation - [Qwen Code Interpreter](https://www.alibabacloud.com/help/zh/model-studio/qwen-code-interpreter) and [Assistant API Code Interpreter](https://www.alibabacloud.com/help/zh/model-studio/code-interpreter) (Last updated: December 2025)
====================Tool Invocation Count====================

Web Extractor invocations: 1

Invocation methods

Web extractor supports three APIs. Use Responses API for new integrations -- it provides the most control.

API	Tool configuration	Streaming required	Notes
Responses API (recommended)	Add `web_search` and `web_extractor` to `tools`	No	Exposes intermediate tool execution status
Chat Completions API	Set `enable_search: true`, `search_strategy: "agent_max"`	Yes	Non-streaming not supported
DashScope API	Set `enable_search: true`, `search_strategy: "agent_max"`	Yes	Java SDK not supported

When using qwen3-max-2026-01-23, set enable_thinking to true.

Responses API

response = client.responses.create(
    model="qwen3-max-2026-01-23",
    input="<your-prompt>",
    tools=[
        {"type": "web_search"},
        {"type": "web_extractor"},
        {"type": "code_interpreter"}
    ],
    extra_body={
        "enable_thinking": True
    }
)

Chat Completions API

completion = client.chat.completions.create(
    model="qwen3-max-2026-01-23",
    messages=[{"role": "user", "content": "<your-prompt>"}],
    extra_body={
        "enable_thinking": True,
        "enable_search": True,
        "search_options": {"search_strategy": "agent_max"}
    },
    stream=True
)

DashScope API

from dashscope import Generation

response = Generation.call(
    model="qwen3-max-2026-01-23",
    messages=[{"role": "user", "content": "<your-prompt>"}],
    enable_search=True,
    search_options={"search_strategy": "agent_max"},
    enable_thinking=True,
    result_format="message",
    stream=True,
    incremental_output=True
)

Stream responses

Web extraction can take time. Stream responses to receive reasoning, tool calls, and output in real time.

Tip: Responses API exposes tool execution status, making it ideal for streaming.

Responses API

Python

import os
from openai import OpenAI

client = OpenAI(
    # If the environment variable is not configured, replace with: api_key="sk-xxx"
    api_key=os.getenv("DASHSCOPE_API_KEY"),
    base_url="https://dashscope.aliyuncs.com/api/v2/apps/protocols/compatible-mode/v1"
)

stream = client.responses.create(
    model="qwen3-max-2026-01-23",
    input="Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it",
    tools=[
        {"type": "web_search"},
        {"type": "web_extractor"},
        {"type": "code_interpreter"}
    ],
    stream=True,
    extra_body={"enable_thinking": True}
)

reasoning_started = False
output_started = False

for chunk in stream:
    # Reasoning process
    if chunk.type == 'response.reasoning_summary_text.delta':
        if not reasoning_started:
            print("=" * 20 + "Reasoning Process" + "=" * 20)
            reasoning_started = True
        print(chunk.delta, end='', flush=True)
    # Tool invocation completion
    elif chunk.type == 'response.output_item.done':
        if hasattr(chunk, 'item') and hasattr(chunk.item, 'type'):
            if chunk.item.type == 'web_extractor_call':
                print("\n" + "=" * 20 + "Tool Invocation" + "=" * 20)
                print(chunk.item.goal)
                print(chunk.item.output)
            elif chunk.item.type == 'reasoning':
                reasoning_started = False
    # Response content
    elif chunk.type == 'response.output_text.delta':
        if not output_started:
            print("\n" + "=" * 20 + "Response" + "=" * 20)
            output_started = True
        print(chunk.delta, end='', flush=True)
    # Response completed -- print tool invocation count
    elif chunk.type == 'response.completed':
        print("\n" + "=" * 20 + "Tool Invocation Count" + "=" * 20)
        usage = chunk.response.usage
        if hasattr(usage, 'x_tools') and usage.x_tools:
            print(f"Web Extractor invocations: {usage.x_tools.get('web_extractor', {}).get('count', 0)}")
            print(f"Web Search invocations: {usage.x_tools.get('web_search', {}).get('count', 0)}")

Node.js

import OpenAI from "openai";
import process from 'process';

const openai = new OpenAI({
    // If the environment variable is not configured, replace with: apiKey: "sk-xxx"
    apiKey: process.env.DASHSCOPE_API_KEY,
    baseURL: "https://dashscope.aliyuncs.com/api/v2/apps/protocols/compatible-mode/v1"
});

async function main() {
    const stream = await openai.responses.create({
        model: "qwen3-max-2026-01-23",
        input: "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it",
        tools: [
            { type: "web_search" },
            { type: "web_extractor" },
            { type: "code_interpreter" }
        ],
        stream: true,
        enable_thinking: true
    });

    let reasoningStarted = false;
    let outputStarted = false;

    for await (const chunk of stream) {
        // Reasoning process
        if (chunk.type === 'response.reasoning_summary_text.delta') {
            if (!reasoningStarted) {
                console.log("====================Reasoning Process====================");
                reasoningStarted = true;
            }
            process.stdout.write(chunk.delta);
        }
        // Tool invocation completion
        else if (chunk.type === 'response.output_item.done') {
            if (chunk.item && chunk.item.type === 'web_extractor_call') {
                console.log("\n" + "====================Tool Invocation====================");
                console.log(chunk.item.goal);
                console.log(chunk.item.output);
            } else if (chunk.item && chunk.item.type === 'reasoning') {
                reasoningStarted = false;
            }
        }
        // Response content
        else if (chunk.type === 'response.output_text.delta') {
            if (!outputStarted) {
                console.log("\n" + "====================Response====================");
                outputStarted = true;
            }
            process.stdout.write(chunk.delta);
        }
        // Response completed -- print tool invocation count
        else if (chunk.type === 'response.completed') {
            console.log("\n" + "====================Tool Invocation Count====================");
            const usage = chunk.response.usage;
            if (usage && usage.x_tools) {
                console.log(`Web Extractor invocations: ${usage.x_tools.web_extractor?.count || 0}`);
                console.log(`Web Search invocations: ${usage.x_tools.web_search?.count || 0}`);
            }
        }
    }
}

main();

curl

curl -X POST https://dashscope.aliyuncs.com/api/v2/apps/protocols/compatible-mode/v1/responses \
-H "Authorization: Bearer $DASHSCOPE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
    "model": "qwen3-max-2026-01-23",
    "input": "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it",
    "tools": [
        {"type": "web_search"},
        {"type": "web_extractor"},
        {"type": "code_interpreter"}
    ],
    "enable_thinking": true,
    "stream": true
}'

Chat Completions API

Python

import os
from openai import OpenAI

client = OpenAI(
    # If the environment variable is not configured, replace with: api_key="sk-xxx"
    api_key=os.getenv("DASHSCOPE_API_KEY"),
    base_url="https://dashscope.aliyuncs.com/compatible-mode/v1"
)

stream = client.chat.completions.create(
    model="qwen3-max-2026-01-23",
    messages=[
        {"role": "user", "content": "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it"}
    ],
    extra_body={
        "enable_search": True,
        "search_options": {"search_strategy": "agent_max"}
    },
    stream=True
)

for chunk in stream:
    print(chunk)

Node.js

import OpenAI from "openai";
import process from 'process';

const openai = new OpenAI({
    // If the environment variable is not configured, replace with: apiKey: "sk-xxx"
    apiKey: process.env.DASHSCOPE_API_KEY,
    baseURL: "https://dashscope.aliyuncs.com/compatible-mode/v1"
});

async function main() {
    const stream = await openai.chat.completions.create({
        model: "qwen3-max-2026-01-23",
        messages: [
            { role: "user", content: "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it" }
        ],
        enable_search: true,
        search_options: { search_strategy: "agent_max" },
        stream: true
    });

    for await (const chunk of stream) {
        console.log(chunk);
    }
}

main();

curl

curl -X POST https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions \
-H "Authorization: Bearer $DASHSCOPE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
    "model": "qwen3-max-2026-01-23",
    "messages": [
        {"role": "user", "content": "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it"}
    ],
    "enable_search": true,
    "search_options": {"search_strategy": "agent_max"},
    "stream": true
}'

DashScope API

The Java SDK is not supported.

Python

import os
import dashscope
from dashscope import Generation

# If the environment variable is not configured, replace with: dashscope.api_key = "sk-xxx"
dashscope.api_key = os.getenv("DASHSCOPE_API_KEY")

response = Generation.call(
    model="qwen3-max-2026-01-23",
    messages=[
        {"role": "user", "content": "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it"}
    ],
    enable_search=True,
    search_options={"search_strategy": "agent_max"},
    enable_thinking=True,
    result_format="message",
    stream=True,
    incremental_output=True
)

reasoning_started = False
output_started = False
last_usage = None

for chunk in response:
    if chunk.status_code == 200:
        message = chunk.output.choices[0].message

        # Reasoning process
        if hasattr(message, 'reasoning_content') and message.reasoning_content:
            if not reasoning_started:
                print("=" * 20 + "Reasoning Process" + "=" * 20)
                reasoning_started = True
            print(message.reasoning_content, end='', flush=True)

        # Response content
        if hasattr(message, 'content') and message.content:
            if not output_started:
                print("\n" + "=" * 20 + "Response" + "=" * 20)
                output_started = True
            print(message.content, end='', flush=True)

        # Save the last usage info
        if hasattr(chunk, 'usage') and chunk.usage:
            last_usage = chunk.usage

# Print tool invocation count
if last_usage:
    print("\n" + "=" * 20 + "Tool Invocation Count" + "=" * 20)
    if hasattr(last_usage, 'plugins') and last_usage.plugins:
        print(f"Web Extractor invocations: {last_usage.plugins.get('web_extractor', {}).get('count', 0)}")

curl

curl -X POST https://dashscope.aliyuncs.com/api/v1/services/aigc/text-generation/generation \
-H "Authorization: Bearer $DASHSCOPE_API_KEY" \
-H "X-DashScope-SSE: enable" \
-H "Content-Type: application/json" \
-d '{
    "model": "qwen3-max-2026-01-23",
    "input": {
        "messages": [
            {
                "role": "user",
                "content": "Please visit the official Alibaba Cloud Model Studio documentation, find the code interpreter topic and summarize it"
            }
        ]
    },
    "parameters": {
        "enable_thinking": true,
        "enable_search": true,
        "search_options": {
            "search_strategy": "agent_max"
        },
        "result_format": "message"
    }
}'

Streaming event types

When streaming with Responses API, these events track extraction and generation progress:

Event type	Description
`response.reasoning_summary_text.delta`	Reasoning text from the model's thinking process
`response.output_item.done`	Tool call completed. Check `item.type` for `web_extractor_call` to get results
`response.output_text.delta`	Response text chunks
`response.completed`	Response complete. The `usage` field contains tool invocation counts

Supported models

International

Model family	Model IDs
Qwen-Max	`qwen3-max`, `qwen3-max-2026-01-23` (thinking mode)
Qwen-Plus	`qwen3.5-plus`, `qwen3.5-plus-2026-02-15`
Qwen-Flash	`qwen3.5-flash`, `qwen3.5-flash-2026-02-23`
Open-source Qwen	`qwen3.5-397b-a17b`, `qwen3.5-122b-a10b`, `qwen3.5-27b`, `qwen3.5-35b-a3b`

Global

Model family	Model IDs
Qwen-Plus	`qwen3.5-plus`, `qwen3.5-plus-2026-02-15`
Qwen-Flash	`qwen3.5-flash`, `qwen3.5-flash-2026-02-23`
Open-source Qwen	`qwen3.5-397b-a17b`, `qwen3.5-122b-a10b`, `qwen3.5-27b`, `qwen3.5-35b-a3b`

Chinese Mainland

Model family	Model IDs
Qwen-Max	`qwen3-max`, `qwen3-max-2026-01-23` (thinking mode)
Qwen-Plus	`qwen3.5-plus`, `qwen3.5-plus-2026-02-15`
Qwen-Flash	`qwen3.5-flash`, `qwen3.5-flash-2026-02-23`
Open-source Qwen	`qwen3.5-397b-a17b`, `qwen3.5-122b-a10b`, `qwen3.5-27b`, `qwen3.5-35b-a3b`

Billing

Web extractor costs:

Component	Details
Model cost	Extracted web content is appended to the prompt, which increases input tokens. Billed at the model's standard token price. See Models for pricing.
Tool cost	Includes charges for both web extractor and web search invocations.

Web search pricing (per 1,000 invocations):

Region	Price
Chinese Mainland and Global	$0.57341
International	$10.00

The web extractor tool itself is free for a limited time.