Skip to main content

Uncensored — This model declares no content filter. SpicyAPI forwards the request without adding a platform content scan; legal and consent duties remain with the customer.Browse every model at this filtering level
minimax/h3/reference-to-video

MiniMax H3 reference-to-video drives a single shot from up to nine numbered reference images, keeping a character, wardrobe and location steady. Clips run 4 to 15 seconds with generated audio at 768p or 2k.

From $0.037/s
Run model

Workbench

These examples were written for this endpoint. Replace the prompt and sample media URLs with your own accessible inputs, then check the current quote and model availability before submitting.

  • Output5s
    character-continuity-named-refs

    Three references, three jobs: <Picture 1> for her face, <Picture 2> the garment, <Picture 3> the room — naming each by number stops them blending into one scene. The `prompt` carries only the action at `duration_seconds` `5` — sitting, crossing her legs, looking up at camera — with the drift written as a direction, not pulled from a clip. Face stays constant while the scene moves.

    • 768p
    • 5s
  • Output5s
    red-dress-theatre-three-angles

    A third angle on a set-up that already exists, which is what references are for: `reference_image_urls` takes up to nine stills, and between them they fix the performer, the garment and the stage so a new camera position does not quietly become a new production. The prompt is then written for this angle only — the walk, the follow spot, what the garment does as she moves — because the rest has been supplied as pictures.

    • 768p
    • 3:4
    • 5s
  • Output7s
    character-continuity

    Three references, three jobs: Picture 1 for her face, Picture 2 the garment, Picture 3 the room — naming each by number stops them blending into one scene. The `prompt` carries only the action at `duration_seconds` `6` — sitting, crossing her legs, looking up at camera — with the drift written as a direction, not pulled from a clip. Face stays constant while the scene moves.

    • 768p
    • 6s
  • Output5s
    voice-and-room

    Two references carry only what a still can show — her face in Picture 1, the spotlight rig in Picture 2 — and the `prompt` adds what a photo can't: a small club, warm reverb, a murmur from the crowd. At `duration_seconds` `5` there's time for one line into the mic before she smiles, turning a pose into a delivered line. Swap the room, the references still anchor face and light.

    • 768p
    • 5s

This example starts at $0.037 per output second. The Playground updates the estimate when you change the settings.