Image to video
MiniMax H3 Image to Video
Upload a photo and get a MiniMax H3 video with sound.
- 4–15 s · 768P / 2K
- No API key
- No install
≈ 44 credits for this video
Credits are used only when a video is generated. If generation fails, they're returned automatically.
Use only content you have the right to submit. Review the Terms, including Acceptable Use, and report safety or abuse issues.
Task
Generation status
Example: an orange-booth rap duo made with MiniMax H3 text-to-video. See the Hotel Lobby AI guide.
Before you start
- 1Name the subject, the action and the camera move, e.g. "slow push-in".
- 2Say where the light comes from and the mood you want.
- 3For image to video, describe only what should move in the photo.
Every H3 video includes sound. To get music, sound effects or dialogue, describe them in your prompt, e.g. “soft piano music, waves on the shore”. If you leave it out, the model picks the sound itself, often just quiet ambience.
How it works
How MiniMax H3 Image to Video Works
1Your image is the first frame
Image to video starts from a picture instead of a blank page. You give the model a first frame, and it predicts how that exact scene would move: the subject, the camera, the light, and the sound. Because the first frame is fixed, the result keeps the look of your image far more closely than a clip generated from words alone. That makes it the right mode when you already have the shot you want, such as a product photo, a character portrait, a landscape, or a piece of concept art, and you need it to come alive.
2Your prompt directs the motion
Your prompt then tells the model what changes over time. A good image-to-video prompt reads like stage directions: what moves first, where the camera goes, how fast things happen, and what the scene sounds like. If you also upload a last frame, the model plans the motion so the clip ends on that image, which is useful for transitions, reveals, and before-and-after shots. The model returns a finished clip with native audio, so you do not have to add a separate soundtrack to get a usable draft.
3One generator, two starting points
This page and the MiniMax H3 video generator on the home page share the same generator. The difference is where you start. Here, the starting point is your image; on the home page, the starting point is a written prompt.

3 steps
How to Turn a Photo into a Video in 3 Steps
- 01
Upload your first frame
Choose image to video in the generator and upload a JPEG, PNG, or WebP image of up to 5 MB. This picture becomes the opening frame of your clip, so pick one that already has the subject, lighting, and composition you want. If you have a target end shot, upload it as the optional last frame.
- 02
Describe the motion
Write a prompt that explains what should happen after the first frame: how the subject moves, how the camera moves, and how the scene changes. You do not need to re-describe everything in the picture. Focus on action, pacing, and mood. Prompts can be up to 7,000 characters.
- 03
Pick length and resolution, then generate
Set a duration between 4 and 15 seconds and choose 768P or 2K. Credits are reserved when you start, the task moves from queued to processing to succeeded, and the finished video with native audio appears in your account. If the task fails, the reserved credits are released.
Settings
MiniMax H3 Image to Video Settings
The controls are deliberately short. Each one maps directly to a field the model accepts, so what you set is what the model receives.
- First frame
- Required. JPEG, PNG, or WebP, up to 5 MB. The video starts from this image.
- Last frame
- Optional. Add a second image when you want the clip to land on a specific end shot.
- Prompt
- Up to 7,000 characters describing motion, camera, and atmosphere.
- Duration
- Any whole number of seconds from 4 to 15.
- Resolution
- 768P for quick drafts, 2K for the sharpest result. 2K costs more credits per second.
- Audio
- Every MiniMax H3 clip is generated with native audio that matches the scene.
TipAspect ratio selection is offered for text to video. For image to video, crop or frame your source picture the way you want the clip to look before you upload it, and keep the subject clear and well lit. Images with one obvious subject and some empty space around it usually give the model the most room to create natural motion.
Prompt ideas
Prompt Examples for Image to Video
Describe motion, camera, and sound rather than repeating what is already in the picture. These examples show the pattern; replace the details with your own scene.
“Slow push-in on the bottle, condensation runs down the glass, soft studio light sweeps from left to right, shallow depth of field, calm ambient music.”
“She turns her head toward the window and smiles, wind moves her hair, the camera drifts slightly to the right, warm afternoon light, quiet city sounds outside.”
“Clouds roll quickly over the mountains, the lake surface ripples, a slow aerial pull-back reveals the whole valley, birds call in the distance.”
“The lanterns flicker on one by one, light snow begins to fall, the camera glides forward through the street, gentle festival music.”
TipA few habits help: name one main action per clip, say how the camera moves (push-in, pan, orbit, static), describe the pace, and mention the sound you expect. If the first result moves too much, shorten the prompt and remove extra actions. If it moves too little, add a clear verb and a camera direction.
Which mode
Image to Video or Text to Video?
Image to video
Use image to video when the look has to stay exactly right: a real product, a character you have already designed, a brand visual, or a photo you took yourself. The first frame anchors the result, and the optional last frame lets you control where it ends.
Text to video
Use text to video when you are still exploring ideas and do not have a picture yet. You describe the whole scene in words, pick one of six aspect ratios (21:9, 16:9, 4:3, 1:1, 3:4, or 9:16), and let the model compose the shot. Both modes share the same durations, resolutions, native audio, and credit pricing, so you can switch between them freely. Text to video is available in the MiniMax H3 video generator.
FAQ
MiniMax H3 Image to Video FAQ
Pricing for credit packs and plans is on the pricing page, and the full list of supported controls is on the features page.
Is MiniMax H3 image to video free?
Accounts created with Google sign-in receive free trial credits, so you can try image to video before buying anything. Email sign-ups do not include free credits. Free generations have a daily cap. After that, generation uses credits from a plan or credit pack, charged per second of video.
How long can an image-to-video clip be?
Between 4 and 15 seconds, in whole seconds. You choose the length before you generate, and longer clips use more credits.
Does the video include sound?
Yes. MiniMax H3 generates native audio together with the picture, so each clip comes with sound that fits the scene.
Do I need ComfyUI, a GPU, or an API key?
No. Everything runs in your browser on this site. There is nothing to install and no API key to manage.
What images can I upload?
JPEG, PNG, or WebP files up to 5 MB each, for the first frame and the optional last frame. Only upload images you own or have rights to. Sexual content, minors, and a real person’s likeness without their consent are not allowed, and prompts and uploads are screened automatically.
What happens to my credits if a generation fails?
Credits are reserved before the task is sent and settled on the server once the video is generated. If the task fails, the reserved credits are released back to your balance. A video that is generated but blocked by our final safety check is still charged; eligible videos can be submitted for manual review within 7 days.
Bring your photo to life
Upload one image, describe the motion, and get a 4–15 second clip with sound.