Maestro and ReFramework FAQ

Answers to common questions about when to use Maestro versus ReFramework, and how to combine both in processes requiring long-running orchestration and queue-based tasks.

Does Maestro replace ReFramework?

• No. Maestro orchestrates end-to-end processes. The ReFramework still supports running transactional UI performers, especially those that are queue-driven.

When do I use one or the other?

• Use Maestro for long-running, multi-actor flows. Use ReFramework for robust, queue-based UI tasks. Use both when a process needs orchestration plus reliable execution.

How do we use Queues in an Agentic Process?

• Have Maestro start jobs or create queue items, let ReFramework performers process them.

How do we keep REFramework "economies of scale" (log in once, do many)?

• Keep using Queues and Robot Performers for batched UI work. Maestro orchestrates around them.

Should ReFramework ever be used with an Agentic process?

• Yes. Use ReFramework for the RPA performers. Maestro orchestrates the end-to-end flow. Only RPA robots should be queue performers.

Can’t we already do everything with ReFramework?

• You can, but you pay an "orchestration tax": custom chains of queues or state tracking, brittle cross-step retries, weak visibility. Maestro provides native process-level control.

Why isn’t Maestro, or an Agentic Process, a queue "performer" (reading directly from a Queue)?

• Maestro is the orchestrator. Queue performers will remain for RPA robot. An Agentic process maps 1:1 to a unit of work, which is effectively the queue item. This implies that the process is the "queue", not a consumer of it. If you need Maestro to drive off work, modify or replace the dispatcher to invoke Maestro directly (Run Job) instead of creating a queue item.

Read more about UiPath Maestro™ and REFrmework.