Skip to main content

Magnific One

The image model with art direction built in: it develops composition, style, lighting and visual treatment from your prompt before rendering.
Magnific One is a text-to-image API that turns a short prompt into an art-directed image. Instead of rendering the prompt literally, it first decides how the image should look — framing, style, light and treatment — and then renders it. You keep the creative control: explore directions with a quick draft, pick the one you want and render it at full quality.

Key capabilities

  • Art direction from a short prompt: describe the subject and the intent; Magnific One develops the composition, style, lighting and visual treatment
  • Two modes in one endpoint: mode: "quality" returns one finished image; mode: "draft" returns a contact sheet of quick variations as 8 or 16 individual images
  • Up to 4k: in quality mode, choose resolution 1k, 2k or 4k and quality medium or high
  • Style references: guide the look with up to 10 reference_images, in both modes, at no extra cost
  • 10 aspect ratios: from 21:9 to 9:16, including 3:1 banners
  • Async processing: every request returns a task ID immediately. Poll the task endpoint or supply webhook_url to be notified on completion

Use cases

  • Campaign key visuals: art-directed images from a one-line brief
  • Concept exploration: a draft sheet of 16 directions in one call, then a quality render of the winner
  • Editorial and social content: consistent visual treatment across a series using the same style references
  • Banners and covers: wide formats such as 21:9 and 3:1

Generate images with Magnific One

Submit a prompt to create a generation task. The API responds with a task ID; collect the result by polling or via webhook.

POST /v1/ai/magnific-one

Create a new image generation task

GET /v1/ai/magnific-one

List all Magnific One tasks with status

GET /v1/ai/magnific-one/{task-id}

Get task status and results by ID

Parameters

Output

When the task is COMPLETED, its generated array holds:
  • quality mode: one image URL
  • draft mode: 8 or 16 image URLs, one per variation, already cut out of the contact sheet

Pricing

  • A quality request is billed once, at the price of its quality and resolution
  • A draft request is billed once for the whole sheet, at the medium / 2k price — never per image, whether you ask for 8 or 16
  • Reference images do not change the price
See the Pricing page for current rates. The endpoint reference pages above are generated from the OpenAPI specification and are the authoritative, complete parameter list.

Frequently Asked Questions

Magnific One is Magnific’s image model with art direction built in. You submit a prompt and receive a task ID immediately; Magnific One develops the direction of the image and renders it. Poll the task endpoint or configure a webhook to receive the result when processing completes.
Use draft to explore. One request returns 8 or 16 variations of your prompt, billed once as a single medium / 2k render. Pick the direction you like and render it in quality mode at the resolution you need.
With tiles: 8 the variations are laid out as a 2x4 sheet, which is twice as wide relative to each image. Some ratios no longer fit and are approximated to the closest supported one: 1:1, 2:3, 3:2 and 3:4 come back exact, the rest may deviate. With tiles: 16 (a 4x4 sheet) every ratio is honoured exactly, so use it when the shape matters.
No. draft always renders its sheet at medium quality and 2k, so its price does not depend on them. They are accepted and ignored.
They guide the style of the result: palette, treatment, mood. You can send up to 10 publicly accessible image URLs, in either mode. They do not change the price.
Rate limits depend on your subscription tier. See Rate Limits for current limits.
A quality request costs the price of its quality and resolution; a draft request costs one medium / 2k render for the whole sheet. See the Pricing page for current rates.

Best practices

  • Draft first, then render: explore with mode: "draft" and tiles: 16, then re-run the chosen prompt in quality mode
  • Use tiles: 16 when the ratio matters: it is the only draft layout that keeps every aspect ratio exact
  • Keep prompts about intent: describe the subject, mood and purpose; let Magnific One decide the composition and light
  • Reuse references for a series: the same reference_images keep a consistent look across many generations
  • Production integration: use webhook_url instead of polling for scalable applications
  • Error handling: implement retry logic with exponential backoff for 503 errors
  • Mystic: Magnific’s photorealistic image generation, with LoRA styles and characters
  • Upscaler Creative: add detail and resolution to a Magnific One render
  • Image Expand: extend an image beyond its original frame