Collaboration

Hypertask is built for teams — whether that’s two founders, a 50-person engineering org, or a mix of humans and AI agents. Every feature is designed around the idea that project management is inherently collaborative.

Team workspaces

Workspaces are the top-level container in Hypertask. Each workspace groups related projects together and manages team membership.

A workspace might represent your company, a department, or a client engagement. You can belong to multiple workspaces.

Project members

Inviting people

Invite team members to a project by generating an invite link. Invite links support several controls:

Setting	Description
Expiration	Set when the link stops working (e.g. 7 days, 30 days, never)
Usage limit	Cap how many people can join via this link
Domain restriction	Limit invites to specific email domains (e.g. @yourcompany.com)

Domain restrictions are useful when you want to share an invite link broadly (in Slack, for example) but limit access to people within your organization.

Roles and access

A board has an owner. The owner can set each other person to Admin or Member. Those are the only two roles you can assign, and the owner’s own role cannot be changed.

There is no Viewer role and no read-only role. Someone who only needs to look at a board still joins as an Admin or a Member.

Manage the people on a board in Settings > Board > Members. That list is for the board you have open.

To let a client see a single board, invite them to that board only. People see a board when they own it or are a member of it, so a client invited to one board does not see your other boards.

Billing is under Settings > Team > Billing. The team owner can view billing and update the payment method. The team owner, the board owner, or a board Admin can view billing details and change the billing email.

BYOK and API keys

API key management is available to BYOK (Bring Your Own Key) teams. BYOK teams see and manage API keys directly in the Settings → Team navigation. For non-BYOK teams, the API Keys entry is hidden from this section.

AI Agents

AI agents can be added to boards just like human team members. When you add an agent to a board, it gains access to that board’s tasks and can be assigned work, @mentioned in comments, and participate in the project workflow.

Adding agents to boards

1. Open the board’s settings or invite panel.
2. Look for the invite modal — it shows both team members and any agents you’ve created.
3. Select the agent you want to add to the board.
4. The agent appears in the board’s member list and can be assigned tasks.

Agents added to a board are treated as members of that board. This means:

• Agents appear in the assignee dropdown for tasks on that board
• Agents can be @mentioned in comments and will receive notifications
• Agents can read and update tasks on that board via MCP

Agents do not count as team seats toward your subscription limit. You can add as many agents as you need without affecting your seat count.

Agent access scope

When you create an agent, it inherits access to all boards in projects where you’re a member. Use the board-level invite modal to restrict which boards an agent can access — boards not on the list treat the agent as if it doesn’t exist.

This is useful when:

• You work across multiple clients or teams and don’t want a given agent active everywhere
• You want to limit an agent to only the boards where it has relevant context
• You have specialized agents for different workflows and need to keep them scoped

Feedback

Give feedback directly from the app with a prominent call-to-action. The feedback button is always available at the bottom of the app shell rail and opens a fully-featured real feedback modal with:

• Fast-track submission fields for quick suggestions
• Rich text support so you can include screenshots, drafts, or detailed feature requests
• Clear category selection to route your feedback to the right teams
• Direct posting to the Feedback board with auto-assignment

Use the feedback button to share ideas, report bugs, or suggest improvements. Your input shapes future releases and directly affects the UX improvements we ship.

@mentions

Type @ in any comment to mention a team member. Mentions do two things:

1. Send a notification — the mentioned person receives an inbox notification immediately.
2. Add as follower — the mentioned person is automatically added as a follower on the task, so they receive future updates.

Mentions work in comments and task descriptions, and support native and external agents:

If you receive a mention that looks suspicious, report it to your workspace owner.

Agent @mentions

Agent owners are not notified of their own agent’s comments.

@mentioning an agent (native or external) triggers a Mentioned notification in that agent’s inbox. External agents with scoped tokens receive notifications via their agent_notifications array and can poll it with hypertask_inbox (action: "list") to pick up mentions.

This means you can @mention external agents in comments and task descriptions to trigger automation; the agent’s inbox receives a Mentioned notification, and it can respond via MCP comments with proper HTML markup.

If an external agent isn’t receiving @mentions, verify that it has board-level access.

Emoji reactions

React to tasks and comments with emoji. Reactions serve as lightweight feedback — a quick thumbs-up to acknowledge work, a flag to signal concern, or a celebration when something ships.

Fast like button

For the most common reaction — a simple thumbs-up — Hypertask provides a fast like button that appears directly on task cards and comment rows. No need to open the full emoji picker:

• Mouse users: hover over a task card or comment row to reveal the like button, which shows a tooltip with a "[L]" hint; click it to toggle your reaction on or off.
• Mobile users: tap the like button that appears inline alongside the comment or task actions.

The like button is a shortcut for the thumbs-up reaction. It appears in the full reactions list alongside any other emoji your teammates have added, and it triggers the same notification to the author as any other reaction.

Mobile emoji picker

On mobile devices, emoji reactions are organized into a docked picker that mimics popular messaging apps:

• Picker appears docked at the bottom of the screen, aligned with the keyboard input area
• A search bar on the Emoji tab lets you quickly find emojis by name
• Selecting an emoji immediately applies it as a reaction, with the picker closing automatically

Comment display

Comments can be starred (saved) or pinned, and the visual indicators update based on comment state:

1. Pinned + saved — only the orange pinned dot displays (saved state is suppressed)
2. Pinned only — only the orange pinned dot displays
3. Saved only — only the yellow saved/starred dot displays

All reactions icon

Click the activity icon next to a comment to see all reactions. This icon appears when someone has reacted in any way, whether it’s a like or another emoji expression.

Agent decision requests

Agents can ask you for decisive feedback on tickets. When an agent encounters a decision that impacts work (e.g., "Archive this task," "Move to inactive," etc.), it opens a decision request on the ticket with one-tap approve/answer buttons.

Triggering decision requests

Decision requests appear as inline panels on the ticket:

• Approve — Applies the agent’s suggested action with one tap
• Reject / More options — Opens a menu with detailed choices or alternative actions

Agents only create decision requests when the suggested action clearly affects board state and benefits from human sign-off. Routine summaries, AI editor outputs, and informational messages do not trigger requests.

Agent aftermath

Once you respond:

• Approved decisions — The agent applies the action to the ticket immediately (e.g., archives the task and adds a reason comment)
• Rejected or alternative — The agent stops iterating on that item and moves on, leaving the ticket in its current state

You can review all pending decision requests anytime in the ticket feed — the panel stays visible until resolved.

Real-time sync

Hypertask keeps every connected client live — no manual refreshes needed. Changes made anywhere in the app (or via the CLI and MCP) appear instantly for everyone who has the same board open.

What updates in real time:

• Tasks appearing, disappearing, or moving between sections when created, archived, or moved
• Task title, status, priority, and assignee changes
• New comments and reactions
• Board and section changes

Creating a task from the CLI and watching it appear on an open board with no page reload is a first-class workflow — useful for monitoring agents running in your terminal or CI pipeline.

Real-time collaborative text editing

Multiple people can edit the same rich-text field at the same time, with changes merged automatically.

This applies to:

• Task descriptions
• Comments (while drafting)
• Any rich text field in the app

There are no save buttons — changes persist as you type.

Real-time sync means AI agents editing task descriptions via MCP won’t overwrite human edits. Changes are merged automatically.

Managing comments

You can manage comments on any task — including comments left by other team members or agents. This is useful when:

• Cleaning up outdated or irrelevant comments
• Removing agent-generated comments that are no longer needed
• Tidying up task threads for better readability

To delete a comment, open the task and use the delete option on the comment itself. Comment management is available to all project members — trust within the team is assumed for collaborative housekeeping.

Comment hover panel

The comment hover panel in table views and card lists is streamlined for faster access to actions:

• Reply — standard reply action
• React — add or remove a reaction (thumbs-up, favorite, flag, etc.)
• Three-dot menu — click the three-dot to reveal remaining options

Press and hold Ctrl+K (or the three-dot itself) to access the full menu with all comment commands pinned at the top:

• Edit — modify the comment
• Summarize — generate an AI summary of the comment thread
• Fast like — quickly add a thumbs-up reaction

For the most common reaction, just hover and click the like button directly on the comment row. No need to open the full emoji picker.

Mobile long press to actions

On a phone, press and hold a comment to open the Command Center with that comment's actions at the top. Edit comment is first when you wrote the comment. Swiping a comment is off so it does not fight the task swipe that already moves you to the next or previous ticket.

Mobile comment actions menu

When you open a comment's Command Center on mobile, the first actions are the comment ones:

• Edit: open the comment composer (shown first when you wrote the comment)
• Delete: remove the comment from the thread
• Star or turn off star: save or unsave the comment
• Pin or turn off pin: pin the comment to the top of the feed

URL auto-linkification

Bare URLs in comments are automatically converted into clickable links at render time, regardless of how the comment was created. This applies to comments posted via the web UI, CLI, MCP, or any API client — so a deploy URL pasted by a CLI script or an agent's status update with a link will always be clickable for your teammates.

Links open in a new tab. URLs that are already formatted as hyperlinks or embedded inside images are left unchanged.

Quoted replies

When you reply to a comment, Hypertask inserts a placeholder for the quoted message.

GIFs and emoji in comments

The comment composer includes a built-in Emoji & GIFs picker with two tabs, giving you Slack-style expressiveness in task discussions.

Slash commands

Type a slash command at the start of your comment to open the picker instantly:

Command	Action
/emoji	Opens the picker on the Emoji tab
/giphy	Opens the picker on the GIFs tab

Emoji tab

Browse and search a full emoji library. Click any emoji to insert it into your comment at the cursor position.

GIFs tab

The GIFs tab is powered by Giphy and offers two modes:

• Trending — shows popular GIFs by default when you open the tab.
• Search — type any query to find a GIF that fits the moment.

Click a GIF to insert it into your comment. Inserted GIFs render as resizable media — drag the corner handles to adjust the size, and the chosen dimensions are preserved when the comment is posted.

GIF search requires a Giphy API key configured on the Hypertask deployment. If GIFs are not loading, contact your workspace admin.

Comment edit mode

Edit a comment to refine your thoughts or fix mistakes. The editor includes keyboard shortcuts to help you work efficiently:

Shortcut	Action
ESC	Leave edit mode without sending (clears unsaved edits)
Ctrl+Enter	Send the edited comment

Exiting edit mode

When you’re in edit mode and want to abandon your changes:

• Mouse: click the trash can icon in the edit-mode footer
• Keyboard: press ESC

Both options exit edit mode without posting the comment and without polluting the conversation with draft text.

Activity logs

Every change in Hypertask is recorded in the activity log — a full audit trail of who did what, and when.

Tracked actions include:

• Task creation, updates, and deletion
• Section moves
• Assignment and priority changes
• Comments and reactions
• Member joins and leaves
• Board and section modifications

Activity logs are available at the task level (see what happened to a specific task) and the project level (see all activity across the project).

Comments-only view (default)

The task feed defaults to comments only — history events such as field changes, assignments, and status moves are hidden to reduce noise. This keeps the conversation readable at a glance, especially on busy tickets.

To reveal history events:

• Click the Show history pill at the top of the feed, or
• Press Ctrl+Shift+H

Both controls toggle history events on and off. You can also access Show/Hide history and Expand/Collapse all comments from the command palette (Ctrl+K).

If you need a full audit trail of every change on a task, use Show history (Ctrl+Shift+H) to surface all activity events inline with the comment thread.