Dashboard reference

Reference for the Dashboard's stat cards, tabs, approval card, modals, and Tasks Management view. For workflows, see the how-to pages (Review pending approvals, Monitor running tasks, Tasks Management).

---

Stat cards (top of the Dashboard)

Stat	Counts	Subtitle
Active Agents	AI Agents configured in your team. Includes running, suspended, scheduled, and idle.	E.g. "2 running, 2 suspended".
Tasks Running	Tasks currently in an active status across all your AI Agents.	Number of batches currently active.
Pending Approvals	Tasks in waiting_for_human. Matches the Pending Approvals tab count.	"N urgent" when any are urgent.
Completed Today	Tasks that reached a terminal status today.	Today's success rate.

---

Tabs on the Dashboard

Tab	Shows
Pending Approvals	Tasks in waiting_for_human that need a human.
Running	Tasks currently in an active status.
Task Batches	State and history of Batch Process runs.

A View All → link next to the tabs opens Tasks Management.

Empty states

Tab	Empty state
Pending Approvals	✓ "All clear — Your agents are running autonomously. Nothing needs your attention."
Running	💤 "Nothing running right now — Start a conversation with one of your agents or wait for a scheduled trigger to fire."
Task Batches	📋 "No batches yet — When an agent needs to run a tool across multiple items in parallel, it creates a batch."

---

Approval card fields

Each card in Pending Approvals shows:

Field	Notes
AI Agent	Bold label at the top of the card.
Tool	Code-style label (e.g. publish_to_wordpress).
URGENT badge	See URGENT rule.
Chain badge	"⚡ Blocking N parent agents" — present when this task is holding up ancestor tasks.
Action description	Short sentence summarising what's about to happen.
Timestamp	Relative (hover for absolute).
Show parameters ▾	Expands to show the exact input payload key by key. Collapse with Hide ▴.

Action buttons

Button	Behaviour
✓ Approve	Approves and executes the tool immediately.
✎ Edit	Opens the task detail modal to edit parameters before approving.
✕ Reject	Opens the dismiss modal to send feedback and reject.
💬 Chat	Opens the conversation this task belongs to.

Resolved states

State	Message
Approved	✓ "Approved — moved to execution queue" (green).
Dismissed	✕ "Dismissed — feedback sent to agent" (red).

---

URGENT rule

A pending task is marked URGENT when both conditions hold:

1. It is in waiting_for_human.
2. At least one of:
   • It blocks ≥ 3 ancestor tasks (three or more parent AI Agents or batches are parked in waiting_for_children because of this task).
   • It has been waiting > 24 hours since it was created.

URGENT is informational — the approval actions are the same. It flags tasks where your delay is costing the most.

---

Task detail modal (from the ✎ Edit button)

Two tabs:

Tab	Contents
Parameters	Action description, optional chain badge, editable form with one field per parameter, Cancel + ✓ Approve with Changes buttons.
Chat	The conversation embed so you can see the surrounding context.

---

Dismiss modal (from the ✕ Reject button)

Field	Value
Title	"Dismiss: {tool}"
Info note	"After dismissal, the agent will receive your feedback and continue the conversation. You can monitor its response in the Conversations tab."
Label	"Feedback for the agent"
Placeholder	"Tell the agent why and what to do instead…"
Buttons	Cancel, ✕ Dismiss with Feedback

---

Recent Activity

Property	Value
Purpose	Chronological log of agent activity: task completions, approvals, errors, and status changes.
Per-entry	Timestamp (hover for absolute), coloured dot (success / approval / info / error), event text.
Empty state	📝 "No activity yet. Once your agents start working, their actions will appear here."

Recent Activity itself has no filter UI — filters live on Tasks Management.

---

Tasks Management — Tasks tab

Column	Shows
Task	Task name or description.
Batch	Batch number if applicable, otherwise —.
Agent	AI Agent (clickable).
Status	Display label: Success, Running, Failed, Needs Approval, Pending.
Time	When the task ran (hover for absolute).
Output	Short preview.
Actions	⏹ stop (if running/pending), 💬 chat.

Filters:

Filter	Values
Agent	All Agents, or any one AI Agent.
Status	All Statuses, Success, Running, Failed, Pending.

Empty state when filters match nothing: "No tasks match your filters. Try adjusting the agent or status filter above."

---

Tasks Management — Batches tab

Column	Shows
Batch	Name + number; ETA below if running.
Agent	AI Agent that started the batch.
Progress	Progress bar with "M / N" counts and a percentage.
Failed	Number of failed rows.
Credits	Total credits consumed by tasks in the batch.
Status	Running, Complete, Failed, Paused, Pending, Cancelled.
Started	Relative (hover for absolute).

Click any row to open the batch modal.

---

Batch modal

Section	Contents
Header	Batch name + number; Agent: {name}; Started: {timestamp}.
Progress	Large M / N; status badge; progress bar.
Stats	✓ {succeeded} succeeded, ✕ {failed} failed, ◦ {pending} pending, ⏱ {eta} when running.
Footer	"Avg. {seconds}s per task · {credits} credits used"
Controls	⏹ stop, ⬇ download CSV.
Results table	Columns: #, Item, Status, Output.

Export

Export	Format	Limit
Batch → ⬇ Download	CSV	10,000 rows per file.

---

Status labels — display vs underlying

The user-friendly labels you see on rows and badges map to the underlying state-machine statuses. Full definitions: Task statuses.

Display label	Underlying status(es)
Success	success
Running	in_progress
Failed	failed
Needs Approval	waiting_for_human
Pending	pending, scheduled, approved
Waiting for approval	waiting_for_human
Waiting for children	waiting_for_children
Approved	approved
Dismissed	dismissed

Batch status labels: Running, Complete, Failed, Paused, Pending, Cancelled.

---

Related

• Dashboard overview — introduction to the Dashboard.
• Tasks Management — the filterable drill-in.
• Task statuses — the canonical state-machine statuses behind every label.
• Human-in-the-Loop overview — the concepts behind pending tasks.