ultrawork

Parallel task execution dispatches multiple independent agents simultaneously to maximize throughput and reduce total runtime. This approach handles batch task delegation where completion is user-managed, rather than handling persistence or state management.

Parallel agent execution fires multiple independent agent calls simultaneously while routing tasks to appropriate model tiers. This lightweight orchestration mechanism maximizes throughput for concurrent data processing without managing state persistence.

Background runs for long operations are supported during concurrent batch task execution. You can dispatch multiple agents simultaneously while optional background processing handles lengthy operations, leaving completion strictly user-managed.

Multi-agent collaboration without state persistence is managed by firing independent agent calls simultaneously with lightweight orchestration. This approach enforces parallel execution and explicit model routing while avoiding any persistence or state management.

Parallel execution avoids persistence and state management because concurrent tasks are independent and user-managed. This design enforces explicit model routing and lightweight orchestration without the overhead of maintaining state across parallel agent calls.

You should not use parallel execution for multi-agent data processing when tasks are dependent on each other or require state management. This approach is strictly for independent tasks where completion is user-managed and no persistence is needed.