- Get started
- Best practices
- Capabilities
- Skills
- CLI
- Examples
- Examples overview
- Recipe: build an agent
- Recipe: automate with RPA
- Recipe: enter data into a web app
- Recipe: extract data from a desktop app
- Recipe: refactor and test a workflow
- Recipe: a queue-based process with REFramework
- Recipe: build a Maestro Flow
- Recipe: coded app and API workflow
- Recipe: verify a release with Test Manager
- Recipe: extract data from documents with IXP
- Advanced
- Help
This recipe builds a transactional process from the Robotic Enterprise Framework (REFramework) template: it pulls items from an Orchestrator queue, processes each one, and handles retries and exceptions the way an enterprise process should. It shows how a coding agent works with UiPath's standard template for resilient, high-volume automation.
Prerequisites
- A coding agent set up for UiPath work, with an authenticated session (see Install and set up).
- An Orchestrator queue with at least 2 sample items, or a request for the agent to create one.
Build the process
-
Open your coding agent in an empty project folder.
-
Describe the task and name the template:
"Create a new RPA project from the REFramework template that processes items from the
Invoicesqueue in Orchestrator. For each item, read the invoice fields, post them to our finance system, set the transaction status, and retry on failure." -
Review the generated project: the Init, Get Transaction Data, Process, and End Process states, the
Configfile, and the built-in logging and exception handling. -
Complete the Process stage with the business logic for each transaction.
-
Set the queue name, retry count, and any assets in the
Configfile rather than hard-coding them. -
Ask the agent to run the Workflow Analyzer and a single test transaction.
-
Confirm items move through the queue with the right status, such as Successful, Failed, or Retried.
-
Confirm that exceptions are logged.
Result: a queue-driven, transactional process built on REFramework, with retries, logging, and exception handling in place.
The generated structure and run commands vary by agent and prompt. Confirm success from the queue item statuses and the logs, not from the agent's "done" message.