Web Scraper

Web Scraper Skill

Overview

Extract structured data from websites responsibly. Choose the right tool based on whether the page is static HTML or requires JavaScript execution, then handle pagination, rate limiting, and data cleaning.

Tool Selection

| Situation | Tool | |-----------|------| | Static HTML, simple pages | requests + beautifulsoup4 | | JS-rendered content (SPAs, React/Vue) | playwright or selenium | | Heavy scraping / crawling | scrapy | | API available | Use the API - always prefer it |

Step-by-Step Process

Step 1: Check for an API First

Before scraping, check:

• robots.txt at site.com/robots.txt
• The site's developer docs for a public API
• Network tab in DevTools - many "scraped" sites already have an XHR/fetch API

If an API exists, use it. It's faster, more reliable, and respectful.

Step 2: Static HTML Scraping

import requests
from bs4 import BeautifulSoup
import time

headers = {
    "User-Agent": "Mozilla/5.0 (compatible; research-bot/1.0; +https://example.com/bot)"
}

response = requests.get("https://example.com/products", headers=headers, timeout=10)
response.raise_for_status()

soup = BeautifulSoup(response.text, "html.parser")

# Find elements by CSS selector
items = soup.select("div.product-card")
for item in items:
    name = item.select_one("h2.product-name").get_text(strip=True)
    price = item.select_one("span.price").get_text(strip=True)
    link = item.select_one("a")["href"]
    print(name, price, link)

Step 3: JavaScript-Rendered Pages

from playwright.sync_api import sync_playwright
import time

with sync_playwright() as p:
    browser = p.chromium.launch(headless=True)
    page = browser.new_page()
    page.goto("https://example.com/spa-page", wait_until="networkidle")

    # Wait for dynamic content
    page.wait_for_selector("div.results")

    # Extract data
    items = page.query_selector_all("div.product-card")
    for item in items:
        name = item.query_selector("h2").inner_text()
        price = item.query_selector("span.price").inner_text()
        print(name, price)

    browser.close()

Step 4: Handle Pagination

import requests
from bs4 import BeautifulSoup

results = []
page = 1

while True:
    url = f"https://example.com/listings?page={page}"
    response = requests.get(url, headers=headers, timeout=10)
    soup = BeautifulSoup(response.text, "html.parser")

    items = soup.select("div.item")
    if not items:
        break  # No more results

    for item in items:
        results.append(item.get_text(strip=True))

    next_btn = soup.select_one("a.next-page")
    if not next_btn:
        break

    page += 1
    time.sleep(1.5)  # Respectful delay between requests

Step 5: Save and Clean the Data

import pandas as pd
import re

df = pd.DataFrame(results)

# Clean price strings
df["price"] = df["price"].str.replace(r"[^\d.]", "", regex=True).astype(float)

# Normalize URLs
df["url"] = df["url"].apply(lambda u: u if u.startswith("http") else f"https://example.com{u}")

df.to_csv("scraped_data.csv", index=False)

Rate Limiting and Politeness

• Always add delays: time.sleep(1) minimum between requests
• Respect Crawl-delay in robots.txt
• Set a descriptive User-Agent with contact info
• Don't send more than 1 request/second without explicit permission

Anti-bot Countermeasures

• 429 Too Many Requests: back off exponentially, add jitter
• Cloudflare / bot detection: use playwright with a real browser profile; avoid headless detection signatures
• IP blocks: rotate residential proxies if permitted by the site's ToS
• CAPTCHAs: do not attempt to bypass - respect the site's access controls

Legal and Ethical Notes

• Only scrape publicly accessible data
• Don't scrape personal data (names, emails, phone numbers) without a legal basis
• Check the site's Terms of Service before scraping at scale
• Don't store or republish copyrighted content

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: Jignesh-Ponamwar
• Source: Jignesh-Ponamwar/skills-mcp
• License: Apache-2.0
• Homepage: https://skills-mcp-jignesh.vercel.app/

Install and usage instructions live in the source repository linked above.