Python Tutorial: Crawling YY Live Videos via API and Saving Locally

In this tutorial we explore a practical Python web‑scraping example that targets the YY live streaming platform. The goal is to retrieve video information through a publicly available API and download the videos to the local file system.

The API endpoint used is https://api-tinyvideo-web.yy.com/home/tinyvideosv2 . By sending a GET request with a custom User‑Agent header, the server returns a JSON payload containing video metadata, including a resurl field that holds the direct video URL.

First, we simulate the request using the requests library:

<code>url = 'https://api-tinyvideo-web.yy.com/home/tinyvideosv2'
headers = {
    'user-agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/84.0.4147.105 Safari/537.36'
}
response = requests.get(url=url, headers=headers)
data = response.json()
</code>

The variable json_data now holds the JSON structure returned by the API. We iterate over the data_list (found under data['data']['data'] ) to extract each video's title and URL:

<code>for d in data_list:
    video_title = str(d['yyNum']) + '.mp4'
    video_url = d['resurl']
    video_content = requests.get(url=video_url, headers=headers).content
    with open('video\\' + video_title, mode='wb') as f:
        f.write(video_content)
        print('保存完成:', video_title)
</code>

The above code saves each video file using Python's built‑in open function in binary write mode.

Because the API returns different data on each call, we can loop the request multiple times to collect more videos:

<code>url = 'https://api-tinyvideo-web.yy.com/home/tinyvideosv2'
headers = {
    'user-agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/84.0.4147.105 Safari/537.36'
}
for _ in range(page + 1):
    response = requests.get(url=url, headers=headers)
    data = response.json()
    data_list = data['data']['data']
    print(data_list)
</code>

A complete, reusable script that combines the request loop, parsing, and saving logic is provided below:

<code>import requests

def fire(page):
    url = 'https://api-tinyvideo-web.yy.com/home/tinyvideosv2'
    headers = {
        'user-agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/84.0.4147.105 Safari/537.36'
    }
    for _ in range(page + 1):
        response = requests.get(url=url, headers=headers)
        data = response.json()
        data_list = data['data']['data']
        print(data_list)
        for d in data_list:
            video_title = str(d['yyNum']) + '.mp4'
            video_url = d['resurl']
            video_content = requests.get(url=video_url, headers=headers).content
            with open('video\\' + video_title, mode='wb') as f:
                f.write(video_content)
                print('保存完成:', video_title)

if __name__ == '__main__':
    fire(10)
</code>

The article concludes by noting that this approach leverages the API for quick video acquisition, and future posts will cover direct HTML page crawling to fetch videos from different hosts and categories.