main{max-width:820px;margin:0 auto;padding:18px}

WHAT IT DOES

This userscript adds a floating Main Content Extractor panel to any webpage that intelligently identifies the page's primary content area. It scans for semantic HTML elements (main, article, section) and uses structural heuristics to find the best content container. The panel displays the page title, word count, estimated reading time, and a list of all links found in the main content. A "Copy Content" button copies the extracted text to your clipboard, and a "Download as TXT" button saves it as a plain text file. The "Focus Mode" toggle hides all non-essential elements (headers, footers, sidebars, ads, widgets) leaving only the main content visible for distraction-free reading. All panel interactions use the BGPanel component, which automatically persists per-site position and collapse state.

WHERE IT RUNS

The script runs on all websites (*://*/*) through the BotGentz extension. It does not require any special permissions beyond those already granted by the extension. It works in any modern browser that supports the BotGentz userscript environment. The panel appears automatically on every page load, ready to assist with content extraction and focused reading.

HOW TO USE

The panel appears as a floating bubble in the bottom-right corner of your browser window. Click the bubble to expand the panel. The "Page Info" section displays the page title, word count, estimated reading time, and the number of links found. The "Links in Content" section lists all links from the main content area. Click "Copy Content" to copy the extracted main content text to your clipboard. Click "Download TXT" to download the content as a plain text file. Toggle "Focus Mode" to hide all non-essential page elements—the panel remains visible—and toggle it again to restore the page. Click "Refresh" to rescan the page for content after dynamic changes.

SEMANTIC CONTENT EXTRACTION WITH HEURISTIC FALLBACK AND FOCUS MODE

The script uses a multi-layered approach to identify the main content. First, it looks for semantic HTML elements: main, article, and elements with role="main". It then checks for common content class names like .content, .post-content, .entry-content, and .article-content. If no clear container is found, it scans all page elements and scores them based on text word count minus a penalty for link density (to avoid navigation areas). The element with the highest score and at least 20 words is selected as the main content. If no suitable element is found, the entire body is used. Once identified, the script extracts clean text by cloning the container, removing script, style, and noscript elements, and normalizing whitespace. Focus Mode works by identifying and hiding elements outside the main content container, including semantic headers, footers, navs, asides, and common sidebar and ad selectors, while preserving the main content and the panel itself.

THE PANEL

The panel is built using the BGPanel component, which provides a consistent, accessible interface across all BotGentz tools. It supports drag-to-reposition via the header, collapse on click, and stores the panel's position and state per site using GM_setValue. The panel includes a close button (✕) to dismiss it temporarily—the panel reappears on next page load. The settings section includes a dropdown for selecting panel position (bottom-right, top-right, bottom-left, top-left), which is saved globally. The panel has a fixed width of 340px and a maximum height of 80vh, with scrolling for content that overflows. The focus mode state is persisted per site using GM_setValue.

PLEASE NOTE

Requires the free BotGentz extension to run. The content extraction uses heuristics that may not work perfectly on all websites—some pages may require manual selection of the content area. The script does not collect, store, or transmit any personal data. All content data is read from the page and displayed locally. Focus Mode hides elements by setting display:none; toggle it off to restore the page to its original state. The reading time estimate assumes an average reading speed of 200 words per minute. Downloaded text files contain only the extracted main content, not the full page HTML.