Domain Data Generator: fills model fields with typed values
Some automations need a value that is not a copy of another field. A category chosen from the text of an email, a next step drafted from a note, a set of fields filled from a free-text description. Domain Data Generator is the agent for that. It takes a prompt and returns data in the exact shape of your model, ready for a workflow step to store.
It reflects how Worknano is built: customisable with AI, focused on core functionality, with the same UI, UX and database across modules. The agent does not need its own storage. It produces typed values for the models you already have.
What it does
- Reads your prompt and returns only the values it describes.
- Matches the response schema supplied by the application, so text stays text, numbers stay numbers and choices stay within the allowed options.
- Keeps the facts you gave it. If your prompt contains a name, a date or an amount, it carries them through.
- Generates only what you asked for. It does not add unrelated details.
- Works for creating a new record and for updating fields on an existing one.
When it helps
- Sorting incoming requests. A workflow reads a new support message and asks the agent for a category and a priority. The step stores both on the record.
- Turning notes into records. After a sales call, you paste notes into a field. The agent proposes the contact details, deal stage and next step as typed values.
- Sample content. You are setting up a new module and want realistic example rows for a demo or training, based on a short description.
How it works with your data
Domain Data Generator has no tools. It does not search your workspace and it does not write to the database. The workflow step passes in the prompt and the schema, receives a structured object, and the step then applies it under the same permissions and rules as any other workflow action. If you run it in a workflow you built, the step decides what happens with the result, and you can add a confirmation or approval step in front of the write.
It runs on a fast model, since the job is a small structured task.
What it does not do
- It does not know the rest of your data. If a value depends on a record elsewhere, pass that information in the prompt.
- It does not validate business rules. The model's own field rules still apply, and invalid values are rejected when saved.
- It is not a source of truth. Generated values are suggestions from your prompt, so test the prompt with a few real examples before you let a workflow run it unattended.
For free-form text, use Content Writer. For building the workflow around the step, see Workflow Creator.
Make it yours
Domain Data Generator is a record in the Agents module. Edit its instructions to be stricter about what it may infer, change the model or temperature (a low value suits data), duplicate it per use case, or start blank.
Start with Build your own agent, read How to edit a page assistant and a reusable agent for the editor steps, or see the bigger picture in AI agents and AI assistants.