Luma AI

Luma Uni-1

Luma Uni-1 is a unified image workflow: submit a prompt by itself to generate an image, or add a source image to edit it with natural-language instructions. Optional reference images can guide subject, composition, or style. Standard is the recommended mode for routine iteration, while Max uses the higher-priced quality endpoint. Text-to-image requests can set an aspect ratio and use web search; those two controls do not apply during editing. Omnipix always requests PNG output.

How to use this model

  1. 1

    Write the target scene or edit as a direct instruction, including composition, lighting, materials, and any required visible text.

  2. 2

    Leave Source Image empty for generation, or upload one source image when existing content must be changed or preserved.

  3. 3

    Add only purposeful references: up to nine for generation or eight additional references for editing.

  4. 4

    Start with Standard; switch to Max when the final asset justifies the higher provider cost.

  5. 5

    For generation, choose an aspect ratio and disable web search if current web context is unnecessary; editing ignores both controls.

Uni-1Available

Luma Uni-1

This single Uni-1 entry selects its operation from the inputs: a prompt without a source image generates a new image, while adding a source image sends an edit request. Standard is the recommended mode and Max is the higher-priced quality option. Generation accepts up to nine references and can use aspect ratio and web search; editing accepts up to eight additional references and does not send those generation-only controls. Every request returns PNG output.

Try now

Properties

  • One catalog entry supports both text-to-image generation and source-image editing.
  • Standard and Max select the native uni-1 and uni-1-max models with different base prices.
  • Output is fixed to PNG; aspect ratio and web search apply only to generation.

Best for

  • Unified text-to-image generation and source-image editing
  • Reference-guided compositions and manga-styled artwork

Tips

  • Use Standard for most work and Max when the final result warrants the higher-quality mode.
  • Web search and aspect ratio apply only when generating without a source image.

How to use this model

  1. 1

    Write a precise creation prompt, or describe the exact edit and what must remain unchanged.

  2. 2

    Leave Source Image empty to generate, or upload one source image to switch to editing.

  3. 3

    Use focused references and stay within the nine-image generation or eight-image editing limit.

  4. 4

    Start with Standard and choose Max only when the higher-quality endpoint is worth the additional cost.

  5. 5

    Set aspect ratio and web search only for generation; they are intentionally omitted from edit requests.

Parameters

Prompt

prompt
Required

Describe the image to create, or explain the changes to make when a source image is provided.

Mode

mode
Optional

Standard is recommended for most requests. Max uses the higher-quality native Luma model at a higher price.

Default: standard

  • Standard

  • Max

Source Image

image_url
Optional

Optional image to edit. Adding it automatically selects the native Luma Uni-1 image-edit operation.

Reference Images

reference_image_urls
Optional

Optional visual references: up to 9 for generation, or up to 8 when editing a source image.

Aspect Ratio

aspect_ratio
Optional

Optional output aspect ratio for text-to-image generation. It is not sent for editing requests.

  • 3:1

  • 2:1

  • 16:9

  • 3:2

  • 1:1

  • 2:3

  • 9:16

  • 1:2

  • 1:3

Style

style
Optional

Use automatic style selection or request a manga-oriented visual treatment.

Default: auto

  • Auto

  • Manga

Enable Web Search

enable_web_search
Optional

Let Luma use web search for text-to-image generation. This setting is not sent when editing a source image.

Default: true
Uni-1Retired

Luma Uni-1

Luma Uni-1 is a unified image model for creating original visuals from text or editing a source image with optional references, selectable Standard and Max quality modes, style control, and generation-time web search.

Properties

  • Supports text to image workflows.
  • Supports image to image workflows.

Best for

  • Unified text-to-image generation and source-image editing
  • Reference-guided compositions and manga-styled artwork

Tips

  • Use Standard for most work and Max when the final result warrants the higher-quality mode.
  • Web search and aspect ratio apply only when generating without a source image.

How to use this model

  1. 1

    Describe the subject, composition, lighting, and style you want.

  2. 2

    Choose an aspect ratio and output size for the intended placement.

  3. 3

    Generate, review the result, then refine the description or add references.

Parameters

Prompt

prompt
Required

Describe the image to create, or explain the changes to make when a source image is provided.

Mode

mode
Optional

Standard is recommended for most requests. Max uses Luma's higher-quality endpoint at a higher price.

Default: standard

  • Standard

  • Max

Source Image

image_url
Optional

Optional image to edit. Adding it automatically selects Luma Uni-1's editing endpoint.

Reference Images

reference_image_urls
Optional

Optional visual references: up to 9 for generation, or up to 8 when editing a source image.

Aspect Ratio

aspect_ratio
Optional

Optional output aspect ratio for text-to-image generation. It is not sent for editing requests.

  • 3:1

  • 2:1

  • 16:9

  • 3:2

  • 1:1

  • 2:3

  • 9:16

  • 1:2

  • 1:3

Style

style
Optional

Use automatic style selection or request a manga-oriented visual treatment.

Default: auto

  • Auto

  • Manga

Enable Web Search

enable_web_search
Optional

Let Luma use web search for text-to-image generation. This setting is not sent when editing a source image.

Default: true