Seedance 2 Character: AI Image Editor

[Beta] Turn fictional character references into reusable video characters. Upload reference images and describe the outfit to get a character_id you can use in SD 2.0 Omni Reference.

📝

Overview

About this model

⚠️ Beta Feature — results may vary. Create a reusable fictional character for video generation. Upload 1–3 reference images of your original character or illustrated artwork, describe the outfit, and get back a character_id you can use in SD 2.0 Omni Reference. This feature is intended for fictional, original, and illustrated characters — not real people or celebrities. Output quality is experimental and not guaranteed.

1Original Characters: Bring your own illustrated or fictional characters into AI-generated videos.
2Character Consistency: Attempt to maintain a consistent character identity across multiple video generations.
3Storytelling: Build a reusable cast of fictional characters for multi-scene video projects.
💰

Pricing & Value

Cost analysis

muapiapp$0.18 per character (beta pricing)

Flat cost per creation. Character ID can be reused unlimited times at no extra charge. Pricing may change when the feature exits beta.

Fal.aiNot available

No equivalent custom character video pipeline.

ReplicateNot available

No equivalent custom character video pipeline.

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

⚙️

Technical Details

Configuration schema

Reference Imagesarray

1–3 reference photos of the character to build the sheet from.

Default Valuehttps://d3adwkbyhxyrtq.cloudfront.net/ai-images/186/956660418681/1b84d57d-0869-40f7-8ceb-85ea38074022.jpg
Character Namestring

Optional label to identify this character.

Default ValueA Hero
📖

Implementation Guide

Developer documentation

How to Use SD 2 Character

⚠️ Beta Notice: This is an experimental feature. Results are not guaranteed to be perfect. This feature is designed for fictional and original characters — using photos of real people or public figures is not supported and may produce poor or unexpected results. Output quality depends on your reference images and prompt. We recommend testing before using in production workflows.

Tips for Best Results

  • Use clear reference images of your original or illustrated character
  • A simple, neutral background helps isolate the character clearly
  • Describe the outfit in specific detail — vague descriptions produce inconsistent results
  • If the first result isn't what you expected, try regenerating with a different reference image or more detailed outfit description

Steps

  1. Upload Reference Images: Provide 1–3 reference images of your fictional character.

  2. Describe the Outfit: Fill in outfit_description with a specific description (e.g., 'red leather jacket, black jeans, white sneakers').

  3. Optionally Name Your Character: Add a character_name for your own reference.

  4. Submit and Poll: Submit to /sd-2-character. Poll /predictions/{request_id}/result until status is completed.

  5. Use the Character ID: Pass the request_id as character_id in /sd-2.0-omni-reference along with your scene prompt to generate a video featuring your character.

Common Questions

Frequently asked

Is this feature production-ready?

This feature is currently in beta. Output quality can vary depending on the reference images and prompt. We recommend testing before using it in critical workflows. Credits are charged per generation regardless of output quality, as the generation process itself runs to completion.

Can I use photos of real people or celebrities?

No. This feature is designed for fictional, original, and illustrated characters only. Using photos of real people, public figures, or celebrities is not supported and may produce poor results. Such use may also raise legal and ethical concerns.

What kind of reference images work best?

Clear, well-lit images of your original or illustrated character with a simple background. The character should be fully visible with no heavy obstructions. Illustrated or stylized art often works well.

Why didn't my character look accurate in the video?

Character accuracy depends on the quality of your reference images and how specifically you describe the outfit. This is an experimental feature and exact likeness is not guaranteed. Try using a clearer reference image or a more detailed outfit description.

How do I use the character_id in video generation?

Pass the `request_id` returned from a completed `sd-2-character` job as the `character_id` field in `/sd-2.0-omni-reference`. Your character will be used in the video generation.

How many reference images should I use?

1 clear image is enough for most cases. You can upload up to 3 images for more detail. Quality and clarity matter more than quantity.

Can I reuse the same character_id multiple times?

Yes. Once created, a character_id can be used in as many video generations as you like at no additional charge.

Are credits refunded if the output quality is poor?

Credits are charged when the generation completes, regardless of output quality. This is consistent with how all AI generation models work — the compute cost is incurred whether or not the result matches your expectations. We recommend testing with a low-stakes prompt first.