Block editor, real-time collaboration, and new integrations

Document Editing

Briefs now use a block-based editor that feels familiar if you've used any modern editing tools. You can collaborate in real time, see AI edits before applying them, and work with images, tables, and structure without leaving the doc.

• Modern block editor — Write and format briefs with slash commands, drag handles, and inline formatting.
• Real-time collaboration — Edit documents at the same time as teammates and see their cursors and changes as they happen.
• AI diff preview — See AI-suggested edits as highlighted diffs and accept or reject them before they're applied.
• Image upload — Paste or drag images into briefs; they're stored in your workspace.
• Block drag-and-drop — Reorder content by dragging blocks using handles that appear on hover.
• Table editing — Create and edit tables with cell selection and alignment controls.
• Markdown export — Export document content as Markdown for use in other tools.
• Undo/redo — Undo and redo changes with keyboard shortcuts, including in collaborative sessions.
• Editor performance — Faster editor load via lazy loading and optimized startup.
• Document heading structure — AI-generated briefs use a clear heading hierarchy (e.g. H2 → H3 → body).
• Editor loading states — Skeleton loading while the editor initializes instead of a blank screen.
• Content width — Document content is centered with a max width for comfortable reading on wide screens.
• Typography and spacing — Consistent line heights, list spacing, and typography for easier reading.
• Dropdown and toolbar styling — Consistent Notion-style dropdowns and formatting toolbars.
• Keyboard navigation — Move through dropdowns and editor options with the keyboard without unexpected jumps.

Chat & Threading

Chat and threads are built for longer, multi-turn conversations. You can branch from any message, get summaries of long threads, and keep sending messages while the AI is still working—all without losing context or cluttering the main thread.

• Nested thread branching — Start a branch from any message to explore a direction without cluttering the main chat.
• Thread summaries — Get AI-generated summaries of long threads so you can catch up quickly.
• Message queueing — Send multiple messages while the AI is working; they're queued and handled in order.
• @Mention Hamster AI — Type @ in chat to mention Hamster AI for focused AI interaction.
• Reply to specific messages — Reply to any message in a thread with a reply button.
• Thread summary sharing — Share thread summaries with teammates via a link.
• Internal reference links — Click links in AI responses to jump to other briefs or pages.
• Reply to brief — Start a conversation that's tied to the brief document.
• Send to parent thread — Post from a child thread back into the parent with context.
• Thread participants — See who's in a thread with avatars and reply counts.
• Tool call replay — Replay AI tool calls in branched threads to re-run edits or actions.
• Auto-generated thread titles — Branched threads get descriptive titles based on the conversation.
• Chat loading indicator — Centered loading state while waiting for AI responses.
• Chat bubble width — Message bubbles use a consistent max width for a cleaner layout.
• Leave thread — Leave threads you no longer need from the thread menu.
• Plan actions in chat — Plan action buttons are hidden on the Chat tab for a cleaner UI.

Integrations & Context

You can connect the tools you already use so the AI has the right context. Notion and Jira are first-class: connect once, and your content stays in sync and available when working on briefs.

• Notion — Connect your Notion workspace so AI can use your pages for context.
• Jira Cloud — Connect Jira so AI can use your project issues when working on briefs.
• Context from connected tools — Content from Notion and Jira stays in sync and available to AI.

AI & Background Processing

AI can run in the background and keep you in the loop. Start a task, switch context, and get clear feedback when the AI is thinking, planning, or executing—and get notified when long-running work is done.

• Background agent tasks — Start AI tasks and keep working elsewhere; they run in the background.
• Visual AI feedback — See when the AI is thinking, planning, or executing (e.g. overlays or stages).
• Cursor agent integration — Connect your Cursor API key to run background coding agents.
• AI processing queues — Multiple AI requests are queued and processed in order under load.
• Agent webhook notifications — Get notified when background agent tasks finish via webhooks.

Mobile

The app is tuned for phones and small screens: correct scaling, stable tabs, and layouts that don't overflow. Notifications, chat, and threading all work from the mobile header and chat view without extra steps.

• Mobile viewport — Pages scale correctly on mobile with no unwanted zoom and readable text.
• Chat tab persistence — Your selected tab (e.g. Chat) stays selected after a refresh on mobile.
• Notifications in header — Open notifications from the mobile header without going to settings.
• Mobile chat layout — Better spacing, padding, and participant display in chat on mobile.
• Thread overflow — Thread branches stay within the viewport and scroll properly on mobile.
• Toolbar collapse — Extra toolbar actions sit behind a chevron on mobile to save space.
• Tooltips and popovers — Tooltips and popovers show correctly on mobile without overflow or z-index issues.
• Tab centering — Tabs and toolbar actions are centered and aligned on mobile.
• Thread participants on mobile — Thread participants show as stacked avatars with an overflow indicator.
• Touch interactions — Touch handling avoids accidental double-tap zoom on buttons and controls.
• Mobile dropdowns — Dropdowns (e.g. "Leave thread") are aligned and work correctly on mobile.
• Mobile padding — Consistent padding and spacing across mobile screens.

UI Polish & Loading

Smaller improvements make the app feel consistent and predictable: clearer separation between panels, stable navigation height, and loading and error states that don't get in the way.

• Panel divider — A clear divider between document and chat panels.
• Loading glyph — The loading glyph is centered in the layout.