Steps Overview

A Flow consists of multiple steps, each doing a specific task — generating objects with an LLM, training a model, creating images, or transforming and combining data. Steps run sequentially, in the order they appear in your Flow’s steps array.

Step execution

When a Flow runs:

Initialize — the user’s input is validated and stored.
Step-by-step execution — each step runs in sequence: Step 0 → Step 1 → Step 2 → …
Results — if a step succeeds, its output is stored under $.results.[stepId].
Referencing outputs — later steps can read prior steps’ results via JSON Path.

⚠

You cannot reference results of a future step. Step 1 can’t read data from Step 3, because Step 3 hasn’t run yet. If you try, you’ll get an error.

If any step fails, the Flow run also fails. A partial result is still stored, indicating where it broke.

Step types

ObjectGenerator

Use an LLM to produce structured JSON objects following a schema.

ModelGenerator

Train or fine-tune a model on user-uploaded images. Outputs a trainingId for image generation.

ImageGenerator

Create images from a base model or trained model with full control over prompts and settings.

ObjectTransformer

Merge, filter, or reformat JSON data using the Flow Language.

Zipper

Bundle images and JSON files from prior steps into a single downloadable .zip.

Linking steps via JSON Path

Each step writes its output to $.results.[stepId]. Subsequent steps can read from it:

{
  id: 'imageGen',
  type: 'ImageGenerator',
  parameters: {
    trainingId: '$.results.modelTrainer.id',
    images: [
      {
        prompt: 'A painting of a cat in the style of $$',
      },
    ],
  },
}

Flow Language in steps

You can use the Flow Language to transform or select fields in any step’s parameters or result fields:

"$.input.someField" — references user input
"$.results.someStep.nestedValue" — references a previous step’s output
@Map, @Extend, @Index, @Format, @Replace — operators for looping, merging, and string manipulation

{
  id: 'transformer1',
  type: 'ObjectTransformer',
  result: {
    expandedData: [
      [
        '@Map($.results.objectGenStep.items, i)',
        {
          '@Extend': '$$',
          index: '@Index(i)',
        },
      ],
    ],
  },
}

This maps each item from objectGenStep.items, merges it into the new object, and adds an index field.

Example

steps: [
  {
    id: 'objectGen',
    type: 'ObjectGenerator',
    parameters: {
      description: 'Generate product descriptions.',
      schema: { type: 'object', properties: { products: { type: 'array', items: { type: 'object', properties: { name: { type: 'string' }, tagline: { type: 'string' } } } } } },
      input: '$.input.category',
    },
  },
  {
    id: 'myModel',
    type: 'ModelGenerator',
    parameters: {
      datasetId: '$.input.photos',
      settings: { baseModel: 'stable-diffusion-v1-5/stable-diffusion-v1-5', epochs: 40 },
    },
  },
  {
    id: 'imageGen',
    type: 'ImageGenerator',
    parameters: {
      trainingId: '$.results.myModel.id',
      baseModel: 'stable-diffusion-v1-5/stable-diffusion-v1-5',
      images: [
        ['@Map($.results.objectGen.products)', { prompt: '$$.tagline', steps: 25 }],
      ],
    },
  },
  {
    id: 'zipEverything',
    type: 'Zipper',
    parameters: {
      sources: ['imageGen', 'objectGen'],
      filename: 'all_results.zip',
    },
  },
],