Data Generator

Overview

Generates a structured dataset from a one-line topic. You define the columns once on the DataGenerator node and get validated rows back instead of prose.

  1. Inputs - topic (nodetool.input.StringInput) and row_count (nodetool.input.IntegerInput, 3 to 30, default 8).
  2. Prompt (nodetool.text.Prompt) - Fills {{TOPIC}} and {{ROW_COUNT}} into the generation instructions.
  3. DataGenerator (nodetool.generators.DataGenerator) - Asks the selected language model for rows that match the columns. The template defines two string columns, name and color, and selects gpt-5-mini. Edit the columns to change the shape.
  4. Output - The dataframe output goes to generated_data. The node also streams each row as a record output.

Demo

Tags

agents

Workflow Diagram

graph TD topic["StringInput (topic)"] count["IntegerInput (row_count)"] prompt["Prompt"] generator["DataGenerator"] out["Output (generated_data)"] topic -->|TOPIC| prompt count -->|ROW_COUNT| prompt prompt --> generator generator -->|dataframe| out
Edit this page on GitHub