Multi-Agent Concurrency

Cooder supports running multiple Agents simultaneously, letting different tasks progress in parallel.

Creating Multiple Agents

In VS Code, each @cooder message can be dispatched to a new Agent for independent execution, or sent as a follow-up instruction to an existing Agent.

In the Cooder app's device view, tap + to manually create a new named Agent (e.g., "Login Module", "Database Migration").

Monitoring Multiple Agents

The device view displays all Agents as cards, each showing:

• Agent name and label
• Current status (Running / Complete / Paused)
• Most recent log entry
• Unread message badge

Tap any card to open that Agent's detail view.

Concurrency Limits

The number of simultaneously running Agents is capped by your plan:

Plan	Concurrent Agents
Free	1
Pro	5
Ultra	Unlimited

When the limit is reached, new tasks queue until an Agent becomes available.

Tips

• Split by module — Break large tasks into independent units (frontend page, backend API, tests) and assign to separate Agents for parallel execution
• Name your Agents — Use descriptive names ("User Module", "Data Layer") to quickly identify sessions in the list
• Avoid file conflicts — Concurrent Agents editing the same file can cause conflicts; plan task boundaries to minimize overlap

See Pricing for plan details.