Improved Firefox compatibility and extraction reliability.

• Uses Firefox-native browser APIs throughout.
• Improved extraction from JavaScript-rendered pages by waiting for content to stabilise.
• Extracts visible text from the main page, accessible frames and open shadow roots.
• Records the final URL when a page redirects.
• Includes the actual Firefox error message when extraction fails.
• Improved background tab loading, extraction and cleanup.

Version 1.21

• Added a configurable delay between processing URLs.
• URLs are now opened and processed sequentially.
• This helps reduce HTTP 429 rate-limit errors on websites that restrict requests made too quickly.
• Improved reliability so one failed URL does not stop the entire batch.
• Processing now continues in the background after it has started.