Use focused generators
Focused generators produce a particular artifact from the code you supply. They support the broader documentation workflow without replacing a structured Space.
| Capability | What to review before using the result |
|---|---|
| Code documentation | Responsibilities, inputs, outputs and examples against the actual source. |
| Tests | Setup requirements, assertions and whether the tests exercise meaningful behavior. |
| Code optimization | Behavior preservation, dependencies and project-specific conventions. |
| Code comments | Accuracy of parameter, return and exception descriptions. |
| UML diagrams | Relationships and flow direction, not just rendering. |
| Language conversion | Target runtime, libraries and behavior differences. |
| OpenAPI/Swagger generation | Endpoint exposure, authentication, schemas and status codes. |
Generate an artifact
Choose the generator matching your task, supply the required code and filename, and complete its options. UML requires a diagram type; language conversion requires source and target languages. For API generation, include enough source to establish the relevant contract.
Start generation only after checking the input. These are AI generation actions and are subject to account permissions and credits.
Validate the output
A generated test or code change is a draft to review and run in your project. A generated diagram or API specification must match the source before you publish it. Do not infer successful compilation or correctness from a successful generation response.
Use Generations to manage the saved result and add useful documentation to a Space. For automation, see the API reference.
Updated