Quick start

This short tutorial will help you start scraping with Crawlee in just a minute or two. For an in-depth understanding of how Crawlee works, check out the Introduction section, which provides a comprehensive step-by-step guide to creating your first scraper.

Choose your crawler

Crawlee offers the following main crawler classes: BeautifulSoupCrawler, ParselCrawler, and PlaywrightCrawler. All crawlers share the same interface, providing maximum flexibility when switching between them.

BeautifulSoupCrawler

The BeautifulSoupCrawler is a plain HTTP crawler that parses HTML using the well-known BeautifulSoup library. It crawls the web using an HTTP client that mimics a browser. This crawler is very fast and efficient but cannot handle JavaScript rendering.

ParselCrawler

The ParselCrawler is similar to the BeautifulSoupCrawler but uses the Parsel library for HTML parsing. Parsel is a lightweight library that provides a CSS selector-based API for extracting data from HTML documents. If you are familiar with the Scrapy framework, you will feel right at home with Parsel. As with the BeautifulSoupCrawler, the ParselCrawler cannot handle JavaScript rendering.

PlaywrightCrawler

The PlaywrightCrawler uses a headless browser controlled by the Playwright library. It can manage Chromium, Firefox, Webkit, and other browsers. Playwright is the successor to the Puppeteer library and is becoming the de facto standard in headless browser automation. If you need a headless browser, choose Playwright.

before you start

Crawlee requires Python 3.9 or later.

Installation

Crawlee is available as the crawlee PyPI package. The core functionality is included in the base package, with additional features available as optional extras to minimize package size and dependencies. To install Crawlee with all features, run the following command:

pip install 'crawlee[all]'

Then, install the Playwright dependencies:

playwright install

Verify that Crawlee is successfully installed:

python -c 'import crawlee; print(crawlee.__version__)'

For detailed installation instructions see the Setting up documentation page.

Crawling

Run the following example to perform a recursive crawl of the Crawlee website using the selected crawler.

BeautifulSoupCrawler

import asyncio

from crawlee.crawlers import BeautifulSoupCrawler, BeautifulSoupCrawlingContext


async def main() -> None:
    # BeautifulSoupCrawler crawls the web using HTTP requests and parses HTML using the BeautifulSoup library.
    crawler = BeautifulSoupCrawler(max_requests_per_crawl=50)

    # Define a request handler to process each crawled page and attach it to the crawler using a decorator.
    @crawler.router.default_handler
    async def request_handler(context: BeautifulSoupCrawlingContext) -> None:
        context.log.info(f'Processing {context.request.url} ...')
        # Extract relevant data from the page context.
        data = {
            'url': context.request.url,
            'title': context.soup.title.string if context.soup.title else None,
        }
        # Store the extracted data.
        await context.push_data(data)
        # Extract links from the current page and add them to the crawling queue.
        await context.enqueue_links()

    # Add first URL to the queue and start the crawl.
    await crawler.run(['https://crawlee.dev'])


if __name__ == '__main__':
    asyncio.run(main())

ParselCrawler

import asyncio

from crawlee.crawlers import ParselCrawler, ParselCrawlingContext


async def main() -> None:
    # ParselCrawler crawls the web using HTTP requests and parses HTML using the Parsel library.
    crawler = ParselCrawler(max_requests_per_crawl=50)

    # Define a request handler to process each crawled page and attach it to the crawler using a decorator.
    @crawler.router.default_handler
    async def request_handler(context: ParselCrawlingContext) -> None:
        context.log.info(f'Processing {context.request.url} ...')
        # Extract relevant data from the page context.
        data = {
            'url': context.request.url,
            'title': context.selector.xpath('//title/text()').get(),
        }
        # Store the extracted data.
        await context.push_data(data)
        # Extract links from the current page and add them to the crawling queue.
        await context.enqueue_links()

    # Add first URL to the queue and start the crawl.
    await crawler.run(['https://crawlee.dev'])


if __name__ == '__main__':
    asyncio.run(main())

PlaywrightCrawler

import asyncio

from crawlee.crawlers import PlaywrightCrawler, PlaywrightCrawlingContext


async def main() -> None:
    # PlaywrightCrawler crawls the web using a headless browser controlled by the Playwright library.
    crawler = PlaywrightCrawler()

    # Define a request handler to process each crawled page and attach it to the crawler using a decorator.
    @crawler.router.default_handler
    async def request_handler(context: PlaywrightCrawlingContext) -> None:
        context.log.info(f'Processing {context.request.url} ...')
        # Extract relevant data from the page context.
        data = {
            'url': context.request.url,
            'title': await context.page.title(),
        }
        # Store the extracted data.
        await context.push_data(data)
        # Extract links from the current page and add them to the crawling queue.
        await context.enqueue_links()

    # Add first URL to the queue and start the crawl.
    await crawler.run(['https://crawlee.dev'])


if __name__ == '__main__':
    asyncio.run(main())

When you run the example, you will see Crawlee automating the data extraction process in your terminal.

Running headful browser

By default, browsers controlled by Playwright run in headless mode (without a visible window). However, you can configure the crawler to run in a headful mode, which is useful during development phase to observe the browser's actions. You can alsoswitch from the default Chromium browser to Firefox or WebKit.

from crawlee.crawlers import PlaywrightCrawler


async def main() -> None:
    crawler = PlaywrightCrawler(
        # Run with a visible browser window.
        headless=False,
        # Switch to the Firefox browser.
        browser_type='firefox',
    )

    # ...

When you run the example code, you'll see an automated browser navigating through the Crawlee website.

Results

By default, Crawlee stores data in the ./storage directory within your current working directory. The results of your crawl will be saved as JSON files under ./storage/datasets/default/.

To view the results, you can use the cat command:

cat ./storage/datasets/default/000000001.json

The JSON file will contain data similar to the following:

{
    "url": "https://crawlee.dev/",
    "title": "Crawlee · Build reliable crawlers. Fast. | Crawlee"
}

tip

If you want to change the storage directory, you can set the CRAWLEE_STORAGE_DIR environment variable to your preferred path.

Examples and further reading

For more examples showcasing various features of Crawlee, visit the Examples section of the documentation. To get a deeper understanding of Crawlee and its components, read the step-by-step Introduction guide.