muapi Logomuapi Logo
muapi Logomuapi Logo

Minimax Hailuo Subject Reference: AI Image Editor

Minimax’s I2I “Subject Reference” model enables you to transform images while preserving the appearance of a subject using a single reference image. Ideal for maintaining character likeness—features, clothing, or expression—across different styles or settings.

Interactive model controls

Minimax’s I2I “Subject Reference” model enables you to transform images while preserving the appearance of a subject using a single reference image. Ideal for maintaining character likeness—features, clothing, or expression—across different styles or settings.

📝

Overview

About this model

Updated Jul 16, 2026

Minimax's Subject Reference model transforms images while preserving the appearance and likeness of a key subject. Provide one reference image showing how your subject should look, then describe the new context or setting, and the model maintains subject identity while applying the transformation. Ideal for keeping character consistency, maintaining a face across style changes, or preserving object identity in new compositions.

Great for fashion designers showing consistent models across different outfits, character artists maintaining likeness through style shifts, or product photographers showing items in varied contexts. At just $0.01 per image, muapi makes this precision subject-consistency tool accessible for batch workflows and rapid iterations.

1Keep a model's face consistent while generating fashion outfit variations
2Maintain a product's appearance across different lighting, background, and environment changes
3Preserve a character's likeness while shifting artistic styles or illustration techniques
4Generate the same person in multiple poses, expressions, and professional scenarios
💰

Pricing & Value

Cost analysis

muapiapp$0.01 per generation

Ultra-affordable per-image pricing. At just $0.01 per generation with no subscription, this is one of the most cost-effective ways to scale subject-consistent image transformations.

** Competitor pricing is estimated based on similar model architectures and usage tiers.

⚙️

Technical Details

Configuration schema

Promptstring

Text prompt describing the image (max 1500 characters).

Default ValueGenerate the same person dressed in Renaissance-style attire, standing in a candlelit castle hall with ornate tapestries and warm low lighting.
Image URLstring

URL of the input image used to generate image.

Default Valuehttps://d3adwkbyhxyrtq.cloudfront.net/ai-images/186/639717681806/ea2ec165-49d9-4212-9ddb-a78a658444c2.jpg
Aspect RatioEnum (8 options)

Aspect ratio of the output image.

Default Value1:1
Number of imagesint

Number of images generated in single request. Each number will charge separately

Default Value1
📖

Implementation Guide

Developer documentation

  1. Prepare a reference image that clearly shows your subject (person, object, character) with good visibility of identifying features. Upload it to a publicly accessible URL. 2. Write a prompt describing the new context, setting, style, or pose (e.g., 'the model in a beach setting wearing a summer dress' or 'the product photographed against a luxury marble background'). 3. Select your preferred aspect_ratio for the output (1:1, 4:3, 16:9, etc.). 4. Set num_images to generate multiple variations if desired. 5. Submit—the model transforms the context while preserving your subject's appearance.
❓

Common Questions

Frequently asked

What is the cost per image?

Subject Reference generation costs $0.01 per image on muapi, making it one of the most cost-effective image transformation tools. Generate dozens of variations for minimal cost.

How well does it preserve subject identity?

The model prioritizes maintaining the subject's visible characteristics—facial features, clothing, color, texture, and distinctive attributes—while changing the surrounding context, pose, or style elements described in your prompt.

Can I change the subject's clothing or pose?

Yes, describe clothing, pose, and styling changes in your prompt. The model will apply those modifications while keeping the core subject identity (face, build, distinguishing features) consistent across variations.