Web Reader Skill

This skill guides the implementation of web page reading and content extraction functionality using the z-ai-web-dev-sdk package, enabling applications to fetch and process web page content programmatically.

Skills Path

Skill Location: {project_path}/skills/web-reader

This skill is located at the above path in your project.

Reference Scripts: Example test scripts are available in the {Skill Location}/scripts/ directory for quick testing and reference. See {Skill Location}/scripts/web-reader.ts for a working example.

Overview

Web Reader allows you to build applications that can extract content from web pages, retrieve article metadata, and process HTML content. The API automatically handles content extraction, providing clean, structured data from any web URL.

IMPORTANT: z-ai-web-dev-sdk MUST be used in backend code only. Never use it in client-side code.

Prerequisites

The z-ai-web-dev-sdk package is already installed. Import it as shown in the examples below.

CLI Usage (For Simple Tasks)

For simple web page content extraction, you can use the z-ai CLI instead of writing code. This is ideal for quick content scraping, testing URLs, or simple automation tasks.

Basic Page Reading

# Extract content from a web page
z-ai function --name "page_reader" --args '{"url": "https://example.com"}'

# Using short options
z-ai function -n page_reader -a '{"url": "https://www.example.com/article"}'

Save Page Content

# Save extracted content to JSON file
z-ai function \
  -n page_reader \
  -a '{"url": "https://news.example.com/article"}' \
  -o page_content.json

# Extract and save blog post
z-ai function \
  -n page_reader \
  -a '{"url": "https://blog.example.com/post/123"}' \
  -o blog_post.json

Common Use Cases

# Extract news article
z-ai function \
  -n page_reader \
  -a '{"url": "https://news.site.com/breaking-news"}' \
  -o news.json

# Read documentation page
z-ai function \
  -n page_reader \
  -a '{"url": "https://docs.example.com/getting-started"}' \
  -o docs.json

# Scrape blog content
z-ai function \
  -n page_reader \
  -a '{"url": "https://techblog.com/ai-trends-2024"}' \
  -o blog.json

# Extract research article
z-ai function \
  -n page_reader \
  -a '{"url": "https://research.org/papers/quantum-computing"}' \
  -o research.json

CLI Parameters

• --name, -n: Required - Function name (use "page_reader")
• --args, -a: Required - JSON arguments object with:
  • url (string, required): The URL of the web page to read
• --output, -o <path>: Optional - Output file path (JSON format)

Response Structure

The CLI returns a JSON object containing:

• title: Page title
• html: Main content HTML
• text: Plain text content
• publish_time: Publication timestamp (if available)
• url: Original URL
• metadata: Additional page metadata

Example Response

{
  "title": "Introduction to Machine Learning",
  "html": "<article><h1>Introduction to Machine Learning</h1><p>Machine learning is...</p></article>",
  "text": "Introduction to Machine Learning\n\nMachine learning is...",
  "publish_time": "2024-01-15T10:30:00Z",
  "url": "https://example.com/ml-intro",
  "metadata": {
    "author": "John Doe",
    "description": "A comprehensive guide to ML"
  }
}

Processing Multiple URLs

# Create a simple script to process multiple URLs
for url in \
  "https://site1.com/article1" \
  "https://site2.com/article2" \
  "https://site3.com/article3"
do
  filename=$(echo $url | md5sum | cut -d' ' -f1)
  z-ai function -n page_reader -a "{\"url\": \"$url\"}" -o "${filename}.json"
done

When to Use CLI vs SDK

Use CLI for:

• Quick content extraction
• Testing URL accessibility
• Simple web scraping tasks
• One-off content retrieval

Use SDK for:

• Batch URL processing with custom logic
• Integration with web applications
• Complex content processing pipelines
• Production applications with error handling

How It Works

The Web Reader uses the page_reader function to:

1. Fetch the web page content
2. Extract main article content and metadata
3. Parse and clean the HTML
4. Return structured data including title, content, and publication time

Basic Web Reading Implementation

Simple Page Reading

import ZAI from 'z-ai-web-dev-sdk';

async function readWebPage(url) {
  try {
    const zai = await ZAI.create();

    const result = await zai.functions.invoke('page_reader', {
      url: url
    });

    console.log('Title:', result.data.title);
    console.log('URL:', result.data.url);
    console.log('Published:', result.data.publishedTime);
    console.log('HTML Content:', result.data.html);
    console.log('Tokens Used:', result.data.usage.tokens);

    return result.data;
  } catch (error) {
    console.error('Page reading failed:', error.message);
    throw error;
  }
}

// Usage
const pageData = await readWebPage('https://example.com/article');
console.log('Page title:', pageData.title);

Extract Article Text Only

import ZAI from 'z-ai-web-dev-sdk';

async function extractArticleText(url) {
  const zai = await ZAI.create();

  const result = await zai.functions.invoke('page_reader', {
    url: url
  });

  // Convert HTML to plain text (basic approach)
  const plainText = result.data.html
    .replace(/<[^>]*>/g, ' ')
    .replace(/\s+/g, ' ')
    .trim();