MiniMax H3 Spicy Image to Video
Animate a starting image with bolder motion, native stereo audio, optional final-frame guidance, 480p/768p output, and 3–15 second durations.
MiniMax H3 Spicy is a high-motion image-to-video API for turning a starting image into bold, expressive video with native stereo audio.Use one Muapi API key to submit a prompt-led animation, optionally guide the final frame, choose 480p or 768p output, and retrieve the completed video asynchronously.
Animate a starting image with bolder motion, native stereo audio, optional final-frame guidance, 480p/768p output, and 3–15 second durations.
MiniMax H3 Spicy Image to Video is a focused API for developers who already have a strong first frame and want more energetic movement from it. The model preserves the supplied image as the visual starting point, then creates a short video with native stereo audio. Describe the action and camera in a prompt, or let the model direct motion from the image alone.
The workflow complements the broader MiniMax H3 API page: use H3 Spicy for bold image animation, standard H3 Image to Video for the open-weights workflow, and H3 Reference to Video when multiple images, clips, or audio references should guide a new scene.
Turn a still image into expressive, high-motion video while retaining its subject and composition as the starting point.
Receive a video with synchronized audio generated in the same workflow, ready for social, product, and concept-video drafts.
Describe the action, camera movement, lighting, pacing, and mood when you need more control over the animation.
Pass `last_image` alongside `image_url` when the shot should interpolate toward a particular final frame.
Use 480p for quicker, lower-cost iteration or 768p when the final clip needs more detail.
Choose a supported duration for short-form content, campaign concepts, character motion, and production tests.
Create bold motion hooks from approved stills for short-form video, reels, and campaign variations.
Animate product photography with camera movement, light changes, and synchronized sound before a full production.
Bring character art, portraits, and concept frames to life with expressive movement and an optional target ending.
Test camera direction and scene energy from a single keyframe before committing to a longer sequence.
Add a single image-to-video node to an automated creative pipeline using Muapi request IDs and webhooks.
Use a clear, publicly accessible image with the subject and composition you want to preserve. Since the output follows the starting image's aspect ratio, choose the frame for the final delivery format before submitting.
Use `prompt` to specify what moves, how the camera moves, and what the mood should be. Concrete verbs such as orbit, track, push in, ripple, turn, and reveal usually give the model a clearer action than a list of visual adjectives.
A slow camera orbit reveals the glass sculpture as sunlight sweeps across its surface; reflections shimmer, the background remains softly out of focus, and the movement feels deliberate and cinematic.
Submit the image-to-video request asynchronously, store the returned request ID, and poll the standard prediction result endpoint or provide a webhook for completion. Muapi handles the common request lifecycle so the same integration can sit beside other image-to-video models.
| Input | Audio | Resolution | Duration | Price |
|---|---|---|---|---|
| Starting image; optional prompt and last frame | Native stereo | 480p / 768p | 3–15 seconds | $0.058–$0.115/sec |
Pricing is per generated second. The published Muapi rate is rounded up to maintain at least a 30% gross margin over the listed model rates: $0.04/sec at 480p and $0.08/sec at 768p.
Submit a starting image, receive a request ID immediately, and poll the standard prediction result endpoint for the completed video.
curl -X POST https://api.muapi.ai/api/v1/minimax-h3-image-to-video-spicy \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"image_url":"https://example.com/first-frame.jpg","prompt":"A bold camera orbit around a glass sculpture as sunlight flashes across the surface","resolution":"480p","duration":5,"seed":-1}'
# Response: {"request_id":"REQUEST_ID"}Pass `image_url` plus an optional motion prompt and output settings to create the clip.
curl https://api.muapi.ai/api/v1/predictions/REQUEST_ID/result \ -H "x-api-key: YOUR_API_KEY" # Poll until status is completed, then read the video URL from the output.
Poll until the task completes, then read the generated video URL from the output.
MiniMax H3 Spicy is an image-to-video model that turns a starting image into bold, expressive video with native stereo audio and optional final-frame guidance.
Only `image_url` is required. `prompt`, `last_image`, `resolution`, `duration`, and `seed` are optional controls for the generated clip.
Yes. The model returns video with native stereo audio as part of the same generation workflow.
Choose 480p or 768p and a duration from 3 through 15 seconds. The default duration is 5 seconds and the default resolution is 480p.
Yes. Set `last_image` together with `image_url` when you want the generated motion to interpolate toward a target final frame.
Muapi charges $0.058 per generated second at 480p and $0.115 per generated second at 768p. Rates are rounded up to maintain at least a 30% gross margin over the listed model rates.
The submit response returns a request ID. Poll the standard prediction result endpoint or pass a webhook URL to receive completion asynchronously.
APIキーを取得して、わずか数分で統合を完了。月額料金なしの従量課金制です。