It has been quite some time since generating illustrations and videos with AI became possible, and today, anyone with an idea can instantly bring their creative vision to life.
However, the quality of AI-generated content often remains a topic of debate.
Unnatural artifacts frequently persist, such as faces and clothing morphing unexpectedly or audio falling out of sync with visual cues.
Amid these challenges, DomoAI rolled out an update to its AI video generation feature, Omni Reference.
The update adds support for two video generation models: Seedance 2.5, which facilitates character consistency, and MiniMax H3, which generates visuals synchronized to audio.
DomoAI Updates Omni Reference, Enabling Video Generation Models Seedance 2.5 and MiniMax H3
Seedance 2.5 is now on DomoAI.
?️ 30 images
? 10 videos
? 10 audio clips
50 references. One generation.
Hand it your references — images, clips, audio — and the character holds. Every shot.
Try one shot for free? https://t.co/NgmTYx0KbV#DomoAI #Seedance pic.twitter.com/N71iy9ByZ6— DomoAI official (@DomoAI_) August 9, 2026
Omni Reference is the video generation capability of DomoAI, a platform for creating AI-powered videos and images.
It allows creators to provide concrete reference assets directly to the AI—capturing character appearances, motions, visual tone, and audio-synchronized performances that text prompts alone cannot fully convey.
With this latest update, Omni Reference now supports video generation models Seedance 2.5 and MiniMax H3.
Maintaining Character Consistency: Seedance 2.5
Seedance 2.5 is a video generation model designed to preserve character facial features, hairstyles, and outfits consistently across scene transitions and changing camera angles.
Each generation supports up to 30 reference images, up to 10 reference videos, and up to 10 reference audio files.
It is ideal for episodic works where the same characters appear continuously.
Characters Moving to Sound: MiniMax H3
MiniMax H3 is a model that generates video by combining visual inputs, such as character images or videos, with auditory inputs like music and voice lines.
It can generate videos where character mouth movements and facial expressions naturally synchronize with uploaded audio.
| AI Video Generation Model Comparison | ||
|---|---|---|
| Feature | Seedance 2.5 | MiniMax H3 |
| Duration per Generation | 4–30 seconds | 4–15 seconds |
| Reference Images | Up to 30 images | Up to 9 images |
| Reference Videos | Up to 10 videos | Up to 3 videos (Up to 3 combined with audio files; max 15 sec total video) |
| Reference Audio | Up to 10 files | Up to 3 files (Up to 3 combined with reference videos) |
| Total Reference Materials | Up to 50 items (30 images, 10 videos, 10 audio files) | Up to 12 items (Up to 9 images; up to 3 combined video and audio files) |
| Output Resolution | 480P or 720P | 768P, Up to 2K |
| Start / End Frames | Supported depending on available generation mode | Supported |
Entirely Web-Based! How to Use
- 1. Access DomoAI and open Omni Reference.
- 2. Upload a character image, and add motion/camera reference videos or music/voice audio as needed.
- 3. Select Seedance 2.5 or MiniMax H3 based on your creative objective.
- 4. Input text prompts describing scenes, character movements, expressions, and camera work, then generate the video.
- 5. Review generated results, fine-tune reference materials or prompt instructions, and regenerate.
Because Omni Reference runs directly in your web browser, you do not need to install any local software.
Elevating Quality Even Further with This Update!
By strategically switching between models—using Seedance 2.5 when you need to maintain character visual consistency, and MiniMax H3 when animating characters to audio—creators can elevate the overall quality of their projects.
Enjoy creating with DomoAI, now capable of producing even higher-quality productions thanks to this update!
For details on DomoAI usage and pricing plans, visit the official DomoAI website.
? MiniMax H3 is now live on DomoAI.
One multimodal workflow. Multiple ways to create.
✍️ Text → Video
?️ Image → Video
? Audio → Video
? Video → Video
With Omni Reference, combine up to 9 images, 3 videos, and 3 audio references in a single creation.More creative… pic.twitter.com/0vKd2W4Guo
— DomoAI official (@DomoAI_) August 18, 2026
