Skip to main content
Pentagon provides two complementary ways to group agents: pods for spatial organization on the canvas, and teams for logical cross-cutting groupings.

​
Pods

A pod is a colored rectangular region drawn on the Office Canvas. Think of it as a rug on the office floor — any agent desk placed within a pod’s bounds automatically belongs to that pod.

​
Creating a Pod

  1. Select the Pod tool (dashed rectangle icon) from the canvas toolbar.
  2. Click and drag on the canvas to draw the pod’s bounds.
  3. Release the mouse, then name the pod and choose a color and emoji.

​
Pod Properties

PropertyDescription
NameDisplay label shown on the canvas
EmojiVisual identifier
ColorBackground tint for the rectangular region
BoundsGrid-snapped origin, width, and height
Parent podPods can nest inside other pods

​
Nesting

Pods can be nested. A pod drawn inside a larger pod becomes a child of that parent. This lets you create hierarchical spatial groupings — for example, a “Backend” pod containing “API” and “Database” sub-pods.

​
Teams

A team is a logical grouping that exists independently of the canvas layout. Teams appear in the sidebar and are identified by a name and emoji.

​
Creating a Team

Click the New Team button in the canvas toolbar, or use the sidebar’s team section. Give the team a name and emoji.

​
Assigning Agents to Teams

Assign agents to a team through the agent’s context menu or the Detail Panel. An agent can belong to one team.

​
Pods vs Teams

PodsTeams
TypeSpatialLogical
MembershipAutomatic — based on desk positionManual — explicitly assigned
Canvas presenceColored rectangle (rug)None — sidebar only
Agent limitOne pod per agent (based on position)One team per agent
NestingPods can nest inside other podsFlat — no nesting
Use caseVisual organization of related agentsCross-cutting concerns, workflows

​
Shared Resources

Both pods and teams have shared files that their member agents can access:
FilePurpose
TASKS.mdShared task list for the group
MEMORY.mdShared context and knowledge
DISCUSSION.mdThreaded notes and coordination between members
These shared files let agents within the same pod or team coordinate without direct communication. An agent can read its pod’s TASKS.md to pick up group-level work, or write to DISCUSSION.md to share findings with teammates.

​
Combining Pods and Teams

Pods and teams work best together. Use pods for spatial arrangement and teams for workflow roles. Example: A full-stack project with three pods and two teams: 
Canvas layout:
┌─────────────────────────────┐
│  Frontend Pod               │
│  ┌───────┐  ┌───────┐      │
│  │ alice │  │  bob  │      │
│  └───────┘  └───────┘      │
└─────────────────────────────┘
┌─────────────────────────────┐
│  Backend Pod                │
│  ┌───────┐  ┌───────┐      │
│  │ carol │  │  dave │      │
│  └───────┘  └───────┘      │
└─────────────────────────────┘
┌─────────────────────────────┐
│  Infra Pod                  │
│  ┌───────┐                  │
│  │  eve  │                  │
│  └───────┘                  │
└─────────────────────────────┘

Teams (sidebar):
  Feature Team: alice, carol
  Review Team: bob, dave, eve
Alice and Carol are spatially separated (different pods) but logically grouped (same team) to collaborate on a feature. Bob, Dave, and Eve are on the Review Team across all three pods.
  • Office Canvas — where pods live
  • Agents — what pods and teams organize
  • Tasks — shared and individual task lists
  • Memory — shared and individual context