You can create collection rules directly from search. Any query that already works in XSaved can now be promoted into a live filter, including tag:research, author:team, and mixed include or exclude rules.

The result is simpler triage. Instead of dragging saves between folders, you define the rule and let the collection keep itself current.

Review queues now remember collapsed groups, selected density, and sort order for the duration of your session. That reduces the amount of interface friction when you move between fast scanning and deeper reading.

We also tightened result rendering so large lists feel more stable while scrolling.

You can now publish a collection as a shared space and invite collaborators with read-only or editor access. Shared spaces preserve filters, notes, and custom ordering, so a curated list behaves the same for every member.

Spaces are built for lightweight distribution. If you already organize work in XSaved, sharing should feel like extending the same object rather than rebuilding it for a team.

Search ranking has been reworked to prioritize recency, repeat opens, and semantic proximity. In practice, this means you can search with shorter phrases and still recover the item you had in mind.

The parser still supports exact syntax when needed, but the default path is now more forgiving for everyday recall.

Long threads now get a tighter summary block at the top of the saved item. We trim duplicate lines, group replies, and preserve the most useful metadata so the page is readable even before you open the source.

This makes saved discussions function more like working notes and less like archived raw output.

Thread results surface the strongest matching segment first, not just the original title. That makes deep references easier to recover when you remember the idea but not the exact source.

You can still jump to the raw entry, but the first screen now optimizes for recognition and speed.