The current prompt for extracting tags from a link (see here) works well but tends to generate slightly different tags for links with similar content.
To improve the quality of the tags and reduce redundancy, I suggest modifying the prompt to include any existing tags. It should instruct the model to reuse these tags if they match the content. This would help maintain consistency and avoid generating too many unique tags for a single link
The current prompt for extracting tags from a link (see here) works well but tends to generate slightly different tags for links with similar content.
To improve the quality of the tags and reduce redundancy, I suggest modifying the prompt to include any existing tags. It should instruct the model to reuse these tags if they match the content. This would help maintain consistency and avoid generating too many unique tags for a single link