Providers and ModelsAmazon

Amazon Nova Canvas

Amazon Nova Canvas
Amazon · Image Generation
POST /v1/images/generations

Image generation and editing model creating and modifying images from text or image inputs, with inpainting, virtual try-on, and style controls.

This model is deprecated and will be retired on 2026-09-30. After that date, requests to this model will fail. Migrate to a successor model before then.

At a glance

FieldValue
Model idamazon-nova-canvas
Input modalitiesText, Image
Output modalitiesImage
Context window-
Weight precision-
Featuresinpainting, background_removal
Native inferenceNo
NewNo
Supported endpointsPOST /v1/images/generations
Deprecation date2026-09-30

Pricing

ChargeSpecRate
Small Standard (≤1024×1024)per image$0.12
Small Premium (≤1024×1024)per image$0.18
Large Standard (≤2048×2048)per image$0.18
Large Premium (≤2048×2048)per image$0.24

Example request

$curl https://api.empiriolabs.ai/v1/images/generations \
>  -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
>  -H 'Content-Type: application/json' \
>  -d '{"model": "amazon-nova-canvas", "prompt": "a misty forest at dawn", "size": "1024x1024"}'

Parameters

ParameterTypeRequiredDefaultDescription
promptstringyes-Text description of the image to generate.
task_typeenumno"TEXT_IMAGE"TEXT_IMAGE | INPAINTING | OUTPAINTING | IMAGE_VARIATION | COLOR_GUIDED_GENERATION | BACKGROUND_REMOVAL | VIRTUAL_TRY_ON. · Allowed: TEXT_IMAGE, IMAGE_VARIATION, INPAINTING, OUTPAINTING, BACKGROUND_REMOVAL, COLOR_GUIDED_GENERATION, VIRTUAL_TRY_ON
qualityenumno"standard"standard | premium. Premium is slower but higher fidelity. · Allowed: standard, premium
negative_promptstringno""Text description of what to avoid in the image.
number_of_imagesnumberno1How many images to generate (1-5). · Range: 1 – 5
aspect_ratioenumno"1:1"Output aspect ratio (1:1, 16:9, 9:16, 4:3, 3:2, etc.). · Allowed: 1:1, 3:4, 2:3, 9:16, 1:2, 1:3, 1:4, 4:3, 3:2, 16:9, 2:1, 3:1, 4:1
size_scaleenumno"small"Output resolution scale. Higher = larger image, slower generation. · Allowed: small, large
cfg_scalenumberno7.5Prompt adherence (1.1-10). Higher = closer to prompt but less creative. · Range: 1 – 20
seednumberno-Random seed for reproducibility. Same seed + prompt = same output.
imagestringno-Source image (URL or base64) for inpainting / outpainting / variation tasks.
reference_imagestringno-Reference image for style or content matching.
mask_imagestringno-Black/white mask defining the region to edit (white = edit, black = keep).
mask_promptstringno-Natural-language description of the region to mask for inpainting.
inpaint_mask_promptstringno-Same as mask_prompt, for inpainting tasks specifically.
tryon_mask_promptstringno-Natural-language description of the garment region for virtual try-on.
similarity_strengthnumberno0.7How closely to match the reference image (0-1). Lower = looser interpretation. · Range: 0.2 – 1
outpainting_modeenumno"DEFAULT"DEFAULT | PRECISE. Precise preserves edges; default is more creative. · Allowed: DEFAULT, PRECISE
color_palettestringno-Predefined color palette name for color-guided generation.
colorsstringno-Comma-separated hex codes (e.g. “#FF0000,#00FF00”) to constrain the palette.
mask_typeenumno"GARMENT"image | prompt. image = use mask_image, prompt = use mask_prompt to derive the mask. · Allowed: GARMENT, PROMPT
garment_classstringno-Garment category for virtual try-on (top, bottom, dress, etc.).
response_formatenumno"url"url (default) returns a signed URL; b64_json returns inline base64. · Allowed: url, b64_json

Notes

Task types

  • TEXT_IMAGE
  • IMAGE_VARIATION
  • INPAINTING
  • OUTPAINTING
  • BACKGROUND_REMOVAL
  • COLOR_GUIDED_GENERATION
  • VIRTUAL_TRY_ON

Mask

  • Image-based or text-prompt-based

Machine-readable schema: GET https://api.empiriolabs.ai/v1/models/amazon-nova-canvas.