- url-metadata package doesn't work in browser due to CORS
- Restore use of fetchReadableContent with Jina AI proxy
- Extract helper functions for cleaner metadata parsing
- Maintain same functionality with proper browser compatibility
- Remove url-metadata dependency (25 packages)
- Start with empty tags field instead of pre-filling 'boris'
- Track if any metadata was successfully extracted (title, description, or tags)
- Only add 'boris' tag if we extracted something automatically
- Makes the boris tag more meaningful and intentional
- User can still manually add tags for URLs without metadata
- Create normalizeTags helper to eliminate duplication
- Handle both string and array tag formats uniformly
- Reduce tag extraction code from 25 lines to 10 lines
- Cleaner, more maintainable code
- Replace manual regex-based HTML parsing with url-metadata package
- Cleaner code with proper handling of OpenGraph, Twitter Cards, and standard meta tags
- Better handling of keywords (supports both string and array formats)
- More reliable extraction across different website structures
- Removes dependency on fetchReadableContent for metadata
- Significantly reduces code complexity (60+ lines to ~20 lines)
- Set 'boris' as default tag in bookmark creation modal
- Extract tags from keywords meta tag (comma/semicolon separated)
- Extract tags from OpenGraph article:tag properties
- Deduplicate and limit to 5 extracted tags
- Prepend 'boris' to any extracted tags
- Only extract tags if user hasn't modified the tags field
- Use ref to prevent re-fetching same URL
- Improves bookmark organization and discoverability
- Extract title with priority: og:title > twitter:title > <title>
- Extract description with priority: og:description > twitter:description > meta description > first <p>
- OpenGraph tags provide better, curated metadata for sharing
- Twitter Card tags as fallback for social media compatibility
- Improved metadata quality for most modern websites
- Automatically fetch page metadata using r.jina.ai proxy
- Debounced (800ms) to avoid API spam while typing
- Only auto-fills if fields are empty (won't overwrite user input)
- Extracts title from page
- Extracts description from meta tag or first paragraph
- Shows spinner indicator while fetching
- Gracefully handles fetch errors (just skips auto-fill)
- Uses existing fetchReadableContent service
- Added tags input field to bookmark modal (comma-separated)
- Updated createWebBookmark to accept tags array
- Tags are added as 't' tags per NIP-B0 spec
- Added published_at tag with current timestamp
- Moved description to content field (per spec, not summary tag)
- d tag now uses URL without scheme (host + path + search + hash)
- Added helper text to explain tag formatting
- Styled form-helper-text for better UX
- Created webBookmarkService for creating web bookmarks
- Added AddBookmarkModal component with URL, title, and description fields
- Added plus button to sidebar header (visible when logged in)
- Modal validates URL format and publishes to relays
- Auto-refreshes bookmarks after creation
- Styled modal with dark theme matching app design
- Follows NIP-B0 spec: URL in 'd' tag, title and summary tags
