Document images, interactive diagrams, and smoother navigation

This update makes documents more visual and structured, while reducing the steps between finding work, editing it, sharing it, and understanding who is involved.

Add and shape images

Upload an image from the editor, paste or drop it into place, and drag it to another position as the document takes shape. Images support small, medium, content, wide, and full layouts, along with cropping, captions, alt text, descriptions, and transcriptions.

You can also copy the visible crop or open the original image. Image access follows the document's existing permissions, and duplicating an image or document creates independent placements instead of coupling later changes. Connected agents can attach images, inspect their originals, and maintain the descriptive metadata that helps future readers understand them.

Build richer structured documents

Mermaid code blocks now open as interactive diagrams by default. Switch between the diagram and its source, zoom and pan through larger visuals, reset the view, and return directly to editable source when a diagram cannot be rendered.

Tables can use content, wide, or full layouts, with resizable columns on supported desktop devices. Their presentation stays with the collaborative document when people or connected agents continue editing. The editor also adds to-do lists, while Markdown updates preserve task states and formatting around inline code.

Browse the library more directly

All Documents now includes top-level collections before the document list. In workspaces with many root collections, the sidebar keeps recently used roots close at hand while preserving the active collection's path.

Collection breadcrumbs now use readable names and link directly to each ancestor. Right-click a sidebar collection for its available actions, or drop documents onto collection rows to move them. On smaller screens, one tap opens an item and an explicit selection mode handles multi-item actions.

Keep work moving while changes complete

Creating a document now opens a focused editor immediately, so you can begin with the title while the new document is prepared. Moves and other reversible actions also appear in the interface as they happen, with drafts, selections, routes, and recovery controls preserved if a request cannot be completed.

See conversations and activity in context

Document and collection rows now show open comment and action-item totals, including rollups for collection branches. Selecting a document's count opens its comments rail directly, and new anchored comments appear against the selected text as soon as they are submitted.

For team owners, the members roster now separates live human presence from connected-agent activity. Open a member's activity details to see the documents involved, distinguish current agent work from recent activity, and retain last-seen context when someone is offline.

Manage sharing where the work lives

Open sharing directly from a collection header and see whether access is already shared or an invitation is pending. Pending guest invitations can now be resent or given a different permission before acceptance.

Sharing with an existing team member now grants access directly when a private or selected-member collection requires it, or confirms when they already have access. Dropping a document into a collection with a broader audience also pauses for confirmation before access expands.

Give connected agents more precise editing tools

Connected agents can now move an entire section, including its child sections, within one document or into another without rebuilding it block by block. When a heading path does not resolve, Quire identifies the failed part and returns nearby section names so the agent can recover with less extra searching.

Agent edits also preserve more of the document around the requested change: task lists and table presentation round-trip through Markdown, formatting can remain wrapped around inline code, and unaffected comment anchors stay attached through structural edits to lists and tables.