Models

Models are the core resources used for training and image generation. A model starts as metadata, can receive training images, can be trained, and can then be used by generation requests.

Typical statuses are new, training, ready, and deleted.

Model lifecycle

Create or update a model with POST /models.
Upload modelInput images for the model.
Trigger training after the model has enough input images.
Poll GET /models/{modelID} about every 5 minutes until the model status is ready.
Use the model ID with image generation requests.

Training normally takes 20 to 25 minutes. Avoid very frequent polling; a 5-minute interval is a good default for integrations.

Create or update a model

Use POST /models to create a new model. Include id when updating an existing model.

Field	Notes
`name`	Human-readable model name.
`age`	Numeric age metadata.
`gender`	`Male` or `Female`.
`ethnicity`	Optional descriptive metadata used by model setup.
`eyeColor`	Optional descriptive metadata used by model setup.
`id`	Include only when updating an existing model.
`userID`	Present in the schema, but normally inferred from your key.

const payload = {
  name: 'MyModel',
  age: 25,
  gender: 'Male',
  ethnicity: 'white',
  eyeColor: 'blue',
}

const response = await fetch(`${BASE_URL}/models`, {
  method: 'POST',
  headers: {
    ...headers,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify(payload),
})

if (!response.ok) {
  throw new Error(await response.text())
}

const body = await response.json()
const model = body.result

payload = {
    "name": "MyModel",
    "age": 25,
    "gender": "Male",
    "ethnicity": "white",
    "eyeColor": "blue",
}

response = requests.post(f"{BASE_URL}/models", headers=headers, json=payload)
response.raise_for_status()

model = response.json()["result"]

Save the returned model ID; it is used for image uploads, training, and trained-model generation.

List models

const response = await fetch(`${BASE_URL}/models`, {
  headers,
})

if (!response.ok) {
  throw new Error(await response.text())
}

const body = await response.json()
const models = body.result

response = requests.get(f"{BASE_URL}/models", headers=headers)
response.raise_for_status()

models = response.json()["result"]

Get a model

const modelId = '<MODEL_ID>'

const response = await fetch(`${BASE_URL}/models/${modelId}`, {
  headers,
})

if (!response.ok) {
  throw new Error(await response.text())
}

const body = await response.json()
const model = body.result

model_id = "<MODEL_ID>"

response = requests.get(f"{BASE_URL}/models/{model_id}", headers=headers)
response.raise_for_status()

model = response.json()["result"]

Check the model status before starting trained-model generation. A trained model should be ready.

Delete a model

Deleting a model also deletes associated images according to the API reference.

const modelId = '<MODEL_ID>'

const response = await fetch(`${BASE_URL}/models/${modelId}`, {
  method: 'DELETE',
  headers,
})

if (!response.ok) {
  throw new Error(await response.text())
}

model_id = "<MODEL_ID>"

response = requests.delete(f"{BASE_URL}/models/{model_id}", headers=headers)
response.raise_for_status()

API reference

Create or update model

Full request body and response schema for POST /models.

List models

Retrieve all models available to the authenticated API key.

Get model

Inspect one model and its current status.

Delete model

Delete a model and its associated images.

Models

Model lifecycle

Create or update a model

List models

Get a model

Delete a model

API reference

Create or update model

List models

Get model

Delete model

On this page