- 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: build a Maestro Case
- Recipe: coded app and API workflow
- Recipe: build a coded function
- Recipe: verify a release with Test Manager
- Recipe: extract data from documents with IXP
- Advanced
- Help
Build, test, and publish a coded function that normalizes an order payload, using a coding agent.
This recipe builds a coded function that normalizes an order payload — trimming whitespace, standardizing IDs, and converting dates to a consistent format. It shows the full loop from your prompt to a function you can call from an automation.
Prerequisites
- The
uipCLI and skills installed, and an authenticated session (see Install and set up). - An empty working folder, or an existing project you want to add the function to.
Build the function
-
Open your coding agent in your working folder.
-
Describe the transform:
"Create a coded function that normalizes this order payload: trim whitespace from every string field, uppercase the order ID, and convert the order date to ISO 8601. Here's a sample payload: [paste sample JSON]."
-
Review the function the agent scaffolds — the input and output contract, and the transform logic.
-
Ask the agent to run the function locally against your sample payload and a second, malformed one, to see how it handles bad input.
-
Ask the agent to publish the function once it behaves as expected.
Result
A published coded function that normalizes an order payload, ready to be called from a Maestro service task, an agent, or a workflow.
Next steps
Wire the function into the automation that needs it — for example, a Maestro Flow that calls it before routing the order for approval. See Coded functions.
The generated contract and file layout depend on your agent, your prompt, and whether you chose Python or TypeScript/JavaScript. Use the transform's correctness against your sample payloads, not the exact files, as your signal of success.