fuse

AI video generator.Your idea.
In motion.

Create with Seedance 2.5, MiniMax H3 official API or open MiniMax H3 BF16. Generate video from text, images or reference media, with model hosting handled for you.

Explore your options

Sign in to create · Generation is billed by model and plan

fuse / THE NEXT SCENEIllustrated shot sequence
A blue-haired character looks back at a companion with a sword on a rooftop at night
An illustrated close-up of a character holding a keycard to a doorSHOT 01
An illustrated shot continuing the sceneSHOT 03

AI video generation.
Text, images and references.

Generate video from a prompt, animate a still image, or guide a shot with reference media. fuse brings these generation modes into one workspace.

Text to video

Describe the scene, action and camera movement. Generate an AI video from your words.

“Rain at night.
The camera follows her
across the rooftop.”

Image to video

Bring a still image into motion. Use a first frame, or first and last frames, to guide the shot.

Reference to video

Guide generation with images, video or audio. Supported media types depend on the model.

ImageVideoAudio

Input modes vary by model. Compare supported models below. Images are creative illustrations.

AI video models.
Three ways to generate.

Seedance 2.5, MiniMax H3 official API, and open MiniMax H3 BF16. Each supports Reference to Video. Compare official APIs by generation cost, or let Fuse handle the machines and setup for the open model.

Official API

Seedance 2.5

Cinematic illustration

From US$0.16/sec480p · Rates vary by output settings

Start from text, or guide generation with image, video and audio references. Use first and last frames to define where a shot begins and ends.

Per clip
4–30 sec
Resolution
480p / 720p / 1080p
Billing
Usage-based
Official API

MiniMax H3

Cinematic illustration

From US$0.12/sec768P · A lower starting cost

Text, first and last frames, and image, video or audio references are also supported. A lower starting cost gives you more room to try different takes.

Per clip
4–15 sec
Resolution
768P / 2K
Billing
Usage-based
Managed by Fuse

Open MiniMax H3

Cinematic illustration

All the setup.
Handled by Fuse.

Fuse sources the machines, installs ComfyUI, configures workflow nodes and loads official BF16 weights. Bring reference images and a prompt to generate video.

Per clip
4–15 sec
Resolution
480P / 768P
Image references
1–9 images
Explore MiniMax H3 and dedicated compute

Starting prices are before tax, at the resolutions shown. The MiniMax H3 official API includes 5 image references; additional images cost US$0.06 each. Illustrations show creative styles. See all options and rates in Plans.

Words. Frames. Sound.
More to create with.

Seedance 2.5 and MiniMax H3 official APIs, directly in fuse. Start with a prompt, define frames or bring reference media into your next shot.

Character image referenceIMAGE REFERENCE
A vivid archive setting referenceSCENE & STYLE
SOUND
Illustrated cinematic shot of two characters on a rooftop
Creative illustration

Text to video

A scene, an action, a camera move. Start from a written prompt.

First & last frames

Define the opening and ending images of your shot.

Image references

Give characters, settings and visual style a clear reference.

Video & audio references

Bring motion and sound into the generation direction.

Both official models offer these modes. Input combinations follow the selected mode; clip lengths and resolutions are listed above.

Open MiniMax H3.
Fuse handles the setup.

From sourcing machines and installing ComfyUI to loading weights, connecting nodes and testing workflows, Fuse takes care of the setup. The open model becomes a tool you can create with.

MANAGED MiniMax H3 BF16fuseComfyUI

  1. Machines & compute

    Source and configure GPUs, storage and the runtime environment.

  2. ComfyUI setup

    Install ComfyUI, its dependencies and the model weights.

  3. Nodes & workflows

    Configure the nodes and connect the MiniMax H3 generation workflow.

  4. Ready to create

    Initialize the environment and verify it with a test generation.

Shared generation or dedicated compute. Create directly in Fuse.

Your focus: the film.
The setup is on us.

You choose the references, prompts and shots.
Fuse handles the MiniMax H3 environment.

MiniMax H3.
Official BF16 weights.

Official BF16 weights for the generation model and text encoder.

BF16 is a 16-bit floating-point format. The generation model and text encoder use official BF16 weights, without lower-bit quantization or distilled few-step checkpoints. fuse prepares the environment they need.

Generation modelBF16
Text encoderBF16
Generation stepsYour choice

Choose 1–100 steps, with 20 as the default. Adjust the settings for your shot.

Your creative pace.
Your way to generate.

A first experiment, a recurring project or a focused production session. Find the option that fits the way you create.

BIWA from HEIST exploring a cinematic idea at a desk

Exploring a new idea.

Start small. Try a few clips and different models.

Consider: Pay-as-you-go generation

Shared MiniMax H3 BF16 from US$0.05/sec at 480P, or choose an official API for the inputs you need.

See generation rates
SUSEKA from HEIST art-directing a product campaign in a studio

Creating, every month.

Social posts, campaigns and stories on a regular schedule.

Consider: Monthly subscription

Choose around your recurring usage and included benefits. Seedance 2.5 and MiniMax H3 official APIs cover text and multimodal inputs.

Compare subscriptions
Mei Hua and White from HEIST arranging storyboard frames together

One story. Many takes.

Short films and multi-shot projects with focused iteration.

Consider: Dedicated MiniMax H3 BF16 compute

US$5.90 for 3 hours. Exclusive RTX 5090 compute, with ComfyUI and the MiniMax H3 workflow prepared for you. No additional per-video charge during the session.

Explore dedicated sessions

Prices are in US dollars before tax. Dedicated time runs continuously after successful deployment and a test video; charging starts at that point. Official APIs are billed separately. See Plans for eligibility and full terms.

Before your
first shot.

A few practical details
about models, compute and creating.

Do I need to install ComfyUI?

No. For open MiniMax H3 generation, fuse prepares the compute environment, deploys ComfyUI, loads the model and sets up its workflow. You prepare your references and prompts, submit generations and collect your videos in fuse. This is a prepared MiniMax H3 BF16 workflow; uploading custom ComfyUI workflows is not supported.

What exactly uses BF16?

The MiniMax H3 generation model and text encoder use the official BF16 weights. BF16 is a 16-bit floating-point format. fuse does not replace these with lower-bit quantized weights or a distilled few-step checkpoint. Available output settings follow the model configuration.

Can I use MiniMax H3 BF16 with text alone or a video reference?

The current MiniMax H3 BF16 workflow uses reference images with a prompt, supporting 1–9 images. For text-only generation, first and last frames, or video and audio references, choose a compatible input mode in Seedance 2.5 or the MiniMax H3 official API.

How does dedicated compute differ from per-generation use?

Per-generation use runs through shared services and is billed according to the model and settings. Dedicated compute gives you an exclusive MiniMax H3 BF16 machine for a continuous time block, with no additional per-video charges during that session. Seedance 2.5 and MiniMax H3 official APIs remain usage-based.

Is video generation free?

AI script text collaboration is free. Image and video generation are billed according to the model, settings and plan; open models also require compute. Sign in to create, and see Plans for current benefits and rates.

Can I work from my own AI assistant?

With fuse CLI, an AI assistant that can run commands can work with your scripts and references, prepare shots, submit jobs and download results. You can continue choosing takes in the browser. Video generation still runs on the selected cloud service.