MiniMax FastH3-Omni

Reactor is proud to introduce FastH3-Omni, our first model collaboration with the FastVideo team. Built on MiniMax’s original H3 model, it brings real-time video and sound to Reactor and puts the characters, the setting, the action, and the atmosphere in your hands.

Start with an example

What happens

Dialogue optional

One spoken line per row, written as Speaker: exact words. Leave blank for no speech.

Camera and style

Soundscape

Background music optional

Reference pictures

0/9

Add a person, object, or setting. Name it to use that name in your scene description.

Add pictures

Audio references

0/3

Use voice or ambience clips between 2 and 15 seconds. Audio references need at least one picture.

Add audio

Clip length

View the prompt sent to the model

Reactor

Choose an example or make your own scene.