Web-Reader (Webpage Reading)

Return format, optional values: markdown, md, text, txt. The default is md.

Whether to enable Simplified Mode. In Simplified Mode, the page content will be trimmed for readability, retaining only readable content. The default is false.

Whether to include images, default is true

Only include elements that match the CSS selector, such as: [“.article_content”]

Parse the page only after these CSS selector elements appear, such as: [“#main”, “.rich_media_content”]

Exclude elements that match CSS selectors, such as: [“#footer”, “.copyright”]

Whether to append all links in the page to the end of the content. The default is false.

Whether to ignore the cache. For the same URL, we usually encrypt and cache it for a short period to speed up access. The default is false. No charges are incurred during the cache period.

Set Cookie, which is useful for pages that require authentication.

Crawling timeout period, in seconds. The default is 120 seconds (2 minutes).

Whether to add the original source for each paragraph when extracting content. The default is false.

The randomness of content generation by the large model during extraction. The value range is 0.0-1.5, with a default of 0.2.

Timeout for extracting topic-related content, in seconds. The default is 120 seconds.

If you only want to focus on content related to a specific topic on the page, set this parameter to use the large model to extract relevant content (Note: This will incur additional large model call fees, priced at 0.008 RMB per 1,000 tokens).