MiniMax H3 Max Image to Video API

minimax/h3-max/image-to-video

480p / 768p / 1080p · 5–15 sec

MiniMax H3 Max Image to Video turns still images into video with native stereo audio, start and end frame control, and lip-synced dialogue. It preserves the source subject's identity and textures while adding fluid, natural motion and camera work.

Input

Prompt658/50000

Image Urls

Drag and drop files, paste from clipboard, or provide a URL. Required. Add 1–2 images in order: starting frame, then optional ending frame.

2/2 files added

Duration

Resolution

Enable Safety Checker

OutputExample

Continue with

MiniMax H3 Max Text to Video MiniMax H3 Max Reference to Video MiniMax H3 Image to Video

Examples

Animate this starting frame into one continuous five-second shot. The adult kyudo archer steadily completes the draw, releases one arrow toward a target off camera, and holds the final follow-through. Maintain the exact face, dark hakama, long bow shape and autumn garden. The bowstring snaps once, the front sleeve settles, and a few maple leaves flutter gently. Camera makes a restrained lateral move while keeping both hands and the bow visible. Natural real-time motion, correct human anatomy, no duplicated arrows or rubber bow. Native bowstring twang and dry leaves, no dialogue, music, subtitles or logos.

Animate the supplied first frame. Inside the same small research submersible, the marine biologist slowly leans toward the round porthole as a translucent bioluminescent jellyfish drifts past outside. Its cool cyan light travels across the scientist's amazed face and the curved glass. Preserve the face, cabin layout and porthole geometry; jellyfish pulses gently with realistic fluid drag. Slow intimate camera push, one continuous five-second shot. Native low machinery hum, a quiet breath and muted water ambience; no implausible underwater speech. No captions, brands, logos or text.

REST API

Quick Start

Make your first request in three steps. Start with the short example, confirm that it returns a task_id, then retrieve the generated video.

Connect to the Vidgo API

Create an API key, keep it on your server, and send it in the Authorization header as a Bearer token.

Endpoint: POSThttps://api.vidgo.ai/api/generate/submit
Authentication: Authorization: Bearer VIDGO_API_KEY

Submit one generation task

Choose a language and run the short example. A successful submission returns a task_id; it does not wait for the video to finish.

Submit request

curl --request POST \
  --url "https://api.vidgo.ai/api/generate/submit" \
  --header "Authorization: Bearer $VIDGO_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "minimax/h3-max/image-to-video",
  "input": {
    "prompt": "Connect the supplied first and last frames in one continuous five-second documentary shot. The same conservator gently peels the thin conservation tissue away from the old fresco, lowering the removed sheet with both hands. Reveal the warm restored pigments gradually; the delicate tissue flexes naturally without melting into the wall. Keep her face, hands, pale linen shirt, chapel and fresco design consistent. Slow subtle camera push toward her pleased expression, warm dust-filled light unchanged. End on the exact supplied final composition with the sheet lowered. Native quiet paper rustling and chapel room tone, no speech or music, no text or logos.",
    "duration": 5,
    "resolution": "768p",
    "enable_safety_checker": true,
    "image_urls": [
      "https://cdn.vidgo.ai/apis/models/minimax/h3-max/image-to-video/v1/01/input-01.png",
      "https://cdn.vidgo.ai/apis/models/minimax/h3-max/image-to-video/v1/01/input-02.png"
    ]
  }
}'

Wait for the result

Use the returned task_id to check the same task until it reaches finished or failed. On success, read data.files[].file_url.

Track status

GET https://api.vidgo.ai/api/generate/status/{task_id}

Poll about every 2 seconds while the status is not_started or running. Stop on finished or failed; on success, read data.files[].file_url. For long-running jobs, increase the interval gradually, or add callback_url to the submit request to receive the final result asynchronously.

not_startedrunningfinishedfailed

Submit response

{
  "code": 200,
  "data": {
    "task_id": "LFE4X1JIHVQ67QET",
    "status": "running",
    "created_time": "2026-10-11T23:00:59"
  }
}

Finished response

{
  "code": 200,
  "data": {
    "task_id": "LFE4X1JIHVQ67QET",
    "status": "finished",
    "files": [
      {
        "file_url": "https://cdn.vidgo.ai/apis/models/minimax/h3-max/image-to-video/v1/01/output.mp4",
        "file_type": "video"
      }
    ],
    "created_time": "2026-10-11T23:00:59",
    "error_message": null,
    "progress": 100
  }
}

Complete runnable example

Open this after the three-step flow is clear. It combines submission, polling, terminal-state handling, and result extraction in one script.

Complete runnable example

set -euo pipefail
: "${VIDGO_API_KEY:?Set VIDGO_API_KEY in your environment}"

REQUEST_BODY=$(cat <<'JSON'
{
  "model": "minimax/h3-max/image-to-video",
  "input": {
    "prompt": "Connect the supplied first and last frames in one continuous five-second documentary shot. The same conservator gently peels the thin conservation tissue away from the old fresco, lowering the removed sheet with both hands. Reveal the warm restored pigments gradually; the delicate tissue flexes naturally without melting into the wall. Keep her face, hands, pale linen shirt, chapel and fresco design consistent. Slow subtle camera push toward her pleased expression, warm dust-filled light unchanged. End on the exact supplied final composition with the sheet lowered. Native quiet paper rustling and chapel room tone, no speech or music, no text or logos.",
    "duration": 5,
    "resolution": "768p",
    "enable_safety_checker": true,
    "image_urls": [
      "https://cdn.vidgo.ai/apis/models/minimax/h3-max/image-to-video/v1/01/input-01.png",
      "https://cdn.vidgo.ai/apis/models/minimax/h3-max/image-to-video/v1/01/input-02.png"
    ]
  }
}
JSON
)

SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --request POST \
  --url "https://api.vidgo.ai/api/generate/submit" \
  --header "Authorization: Bearer $VIDGO_API_KEY" \
  --header "Content-Type: application/json" \
  --data "$REQUEST_BODY")

TASK_ID=$(printf '%s' "$SUBMIT_RESPONSE" | jq -r '.data.task_id // .task_id // empty')
if [ -z "$TASK_ID" ]; then
  printf 'Submit response did not include task_id:
%s
' "$SUBMIT_RESPONSE" >&2
  exit 1
fi

while true; do
  STATUS_RESPONSE=$(curl --silent --show-error --fail-with-body \
    --url "https://api.vidgo.ai/api/generate/status/$TASK_ID" \
    --header "Authorization: Bearer $VIDGO_API_KEY")
  STATUS=$(printf '%s' "$STATUS_RESPONSE" | jq -r '.data.status // .status // empty')

  case "$STATUS" in
    finished)
      printf '%s' "$STATUS_RESPONSE" | jq -r '(.data.files // .files // [])[]?.file_url'
      break
      ;;
    failed)
      printf '%s' "$STATUS_RESPONSE" | jq -r '.data.error_message // .error_message // "Generation failed"' >&2
      exit 1
      ;;
    not_started|running)
      sleep 2
      ;;
    *)
      printf 'Unexpected task status: %s
' "$STATUS" >&2
      exit 1
      ;;
  esac
done

Request Parameters

model and callback_url belong at the request root; generation parameters belong inside input.

Field	Type	Required	Default	Description
model	string	Yes	—	Fixed to minimax/h3-max/image-to-video.
callback_url	string (URL)	No	—	A public HTTP(S) URL for receiving task completion notifications.
input	object	Yes	—	An object containing the parameters required for generation.
input.prompt	string	Yes	—	Describe action development, camera movement, and dialogue. The original input must not exceed 50,000 characters, including leading and trailing whitespace, and must not be empty after trimming.
input.duration	integer	No	5	Generation duration: any whole number from 5–15 seconds.
input.resolution	string	No	768p	Video resolution: 480p, 768p, or 1080p.
input.enable_safety_checker	boolean	No	true	Whether to enable the generated-content safety checker.
input.image_urls	string[]	Yes	—	An image array containing 1–2 public HTTP(S) URLs: the first image is the opening frame, and the optional second image is the ending frame.

Response Fields

Submission returns the task identity immediately. Status responses add progress, output files, or a failure message as the task advances.

Field	Type	Description
code	integer	Application result code. Successful responses use 0 or 200.
message	string	Human-readable response or error message when present.
data.task_id	string	Task identifier used in the status endpoint path.
data.status	string	Current state: not_started, running, finished, or failed.
data.created_time	string	Task creation time in date-time format.
data.progress	number	Completion percentage when the provider reports progress.
data.files[]	array	Generated output files returned after a successful task.
data.files[].file_url	string	Public URL of the generated video.
data.files[].file_type	string	Generated file type, such as video.
data.files[].watermark_url	string \| null	Watermarked output URL when one is available.
data.error_message	string \| null	Failure detail when status is failed.

Task Lifecycle

Treat not_started and running as non-terminal states. finished and failed are terminal alternatives; stop polling when either is returned.

not_started
The task has been accepted and is queued to start.
running
The video is being generated. Continue polling the status endpoint.
finished
The task is complete. Read the video URL from data.files[].file_url.
failed
Generation failed. Read data.error_message for details and stop polling.

Polling and Errors

Authentication401 indicates a missing or invalid Bearer API Key. Check your key in the dashboard before retrying.
Validation400 indicates invalid request parameters. Check the image_urls count and prompt length.
Polling intervalPoll the status every 2 seconds. For longer tasks, use exponential backoff to gradually increase the interval to 5 seconds.
Terminal statesPoll only while the status is not_started or running. Stop immediately on finished or failed.
Callback optionProvide callback_url to receive an automatic notification when generation finishes, without continuous polling.

Model Specifications

Specification	Value	Description
Model	minimax/h3-max/image-to-video	Model identifier for asynchronous image-to-video generation.
Duration	5–15	Whole seconds; defaults to 5 seconds.
Resolution	480p \| 768p \| 1080p	Multiple output resolution levels; defaults to 768p.
Pricing	10 / 16 / 32 credits/s	Corresponds to the 480p / 768p / 1080p tiers, respectively.

View the MiniMax H3 Max Image to Video API documentation

MiniMax H3 Max Image to Video

MiniMax H3 Max Image to Video is an image-to-video model built for faithful animation of still images. Generate coherent 5–15 second videos from a first frame, or use start and end frames to control both ends of the shot. It preserves subject details while generating precisely synchronized lip movements, native dialogue, and ambient sound. With output from 480p to 1080p, it offers detailed visual control for motion comics, e-commerce ads, and film visual effects.

Why choose the MiniMax H3 Max Image to Video API?

Millisecond startup from a first frameUse a single high-quality still as the visual reference to quickly generate coherent motion while preserving facial features, clothing details, material textures, lighting, and composition.
Control both start and end framesProvide opening and closing keyframes to precisely guide how the action develops and where the camera movement ends, keeping transitions and final poses under control.
Precise dialogue and lip syncFor close-ups or medium shots, add dialogue to the prompt to generate clear speech and natural lip movements that match the delivery in a single pass.
High-throughput rendering faster than real timeEnd-to-end inference optimization reduces image-to-video wait times for production pipelines, supporting concurrent batch creation and creative A/B testing.
Resolution tiers with per-second billingChoose any whole-second duration from 5–15 seconds and 480p/768p/1080p resolution. Billing follows the requested duration parameter, with no hidden premiums.

Parameters

Parameter	Requirement	Description
prompt	Required	Describe the subject's actions from the first frame, camera movement, environmental changes, and sound or dialogue cues. The original input must not exceed 50,000 characters, including leading and trailing whitespace, and must not be empty after trimming. Default`—`
duration	Optional	Generation duration must be a whole number from 5–15 seconds. Defaults to 5 seconds. Default`5`
resolution	Optional	Output resolution: 480p, 768p (default), or 1080p. Default`768p`
enable_safety_checker	Optional	Whether to enable the generated-content safety checker to filter sensitive or prohibited content. Default`true`
image_urls	Required	A list of public HTTP(S) image URLs. One image serves as the opening frame; with two images, the first is the opening frame and the second is the ending frame. Supports JPEG, PNG, or WebP. Default`—`

How to Use

Upload the first frameChoose a JPEG, PNG, or WebP image with a clearly defined subject and lighting as the visual starting point for the video.
Add an optional end frameTo end the shot on a specific action or composition, append a second image as the end frame.
Describe actions and dialogueExplain how the subject moves from the first frame and how the camera moves. For speaking characters, specify the English dialogue and lip-sync requirements.
Set duration and resolutionChoose a whole-second duration from 5–15 seconds and 480p/768p/1080p resolution to match your delivery requirements.
Submit asynchronously and retrieve the videoSend an API request to receive a task ID, then poll the status endpoint or configure a callback URL to receive the generated MP4 file.

Usage and billing

The MiniMax H3 Max Image to Video API bills per second. Adding an end frame incurs no extra parameter fee; cost depends only on the requested duration parameter and selected resolution.

Usage	Rate	Details
480p	10 credits / second	$0.050 / second. A 5-second video costs 50 credits; 10 seconds costs 100 credits. Suitable for frequent creative drafts and quick validation.
768p	16 credits / second	$0.080 / second. A 5-second video costs 80 credits; 10 seconds costs 160 credits. Balances throughput and visual detail.
1080p	32 credits / second	$0.160 / second. A 5-second video costs 160 credits; 10 seconds costs 320 credits. Suitable for final commercial delivery and large-screen viewing.

Best Use Cases

E-commerce product and model videosTurn static product shots on white backgrounds or studio model photos into showcase videos with motion and shifting light to improve conversions.
Motion comics and subtle character expressionsAnimate hair and clothing in anime illustrations or character art, or have characters speak specific lines of dialogue.
Film concept development and animated storyboardsTurn keyframe sketches into moving shot previews to assess transitions and character blocking before production.
Bring historical photos to lifeAdd subtle, realistic movement and ambient sound to historical figures or vintage street scenes in old photographs.

Pro Tips

Keep the start and end frames consistent: use the same aspect ratio for both images, and keep the subject's body proportions and lighting as consistent as possible.
Separate actions from camera movement: describe specific gestures such as turning the head, raising a hand, or smiling, then specify camera push-ins, pullbacks, or tracking paths separately.
Use dialogue to strengthen the scene: for lip-synced speech, add a short English line such as "He softly says, 'Welcome home.'" Make sure the character's face and lips are clearly visible in the first image.
Respect the source image's lighting: match the prompt's atmosphere to the image's existing color temperature and shadow direction. Conflicting lighting instructions can cause flicker.

Notes

The MiniMax H3 Max Image to Video API accepts only 1–2 image URLs. Intermediate frames, reference video, and reference audio are not supported.
All images must use HTTP(S) links that the server can access publicly. Private local file paths and direct media-file inputs are not supported.
Submission immediately returns a task_id while the task queues and generates asynchronously. Retrieve the complete video once generation finishes.
Set callback_url at the request root to receive an automatic Webhook notification when the task finishes or fails.

Related Models

Gemini Omni 1.1 Flash Text-to-Video preview

Gemini Omni 1.1 Flash Text-to-VideoGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

Gemini Omni 1.1 Flash Image-to-Video preview

Gemini Omni 1.1 Flash Image-to-VideoGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

Gemini Omni 1.1 Flash Reference-to-Video preview

Gemini Omni 1.1 Flash Reference-to-VideoGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

Gemini Omni 1.1 Flash Video Edit preview

Gemini Omni 1.1 Flash Video EditGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

MiniMax H3 Max Turbo Text to Video preview

MiniMax H3 Max Turbo Text to VideoMiniMax H3 Max Turbo Text to Video turns text into coherent 5–15 second videos with synchronized native stereo audio and fast generation. It follows physical motion while rendering subtle changes in light and shadow.

MiniMax H3 Max Turbo Image to Video preview

MiniMax H3 Max Turbo Image to VideoMiniMax H3 Max Turbo Text to Video turns text into coherent 5–15 second videos with synchronized native stereo audio and fast generation. It follows physical motion while rendering subtle changes in light and shadow.

MiniMax H3 Max Text to VideoMiniMax H3 Max Text to Video turns text prompts into cinematic video with native stereo audio, faster-than-real-time generation, and precise instruction following. It renders complex camera work and physical motion with fine detail across multiple resolutions and common aspect ratios.

MiniMax H3 Max Reference to Video preview

MiniMax H3 Max Reference to VideoMiniMax H3 Max Text to Video turns text prompts into cinematic video with native stereo audio, faster-than-real-time generation, and precise instruction following. It renders complex camera work and physical motion with fine detail across multiple resolutions and common aspect ratios.

Wan 3.0 Prime Text-to-VideoAlibaba Wan 3.0 Prime video model with text-to-video, image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds with optional synchronized audio.

Wan 3.0 Prime Image-to-VideoAlibaba Wan 3.0 Prime video model with text-to-video, image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds with optional synchronized audio.

Wan 3.0 Prime Reference-to-Video preview

Wan 3.0 Prime Reference-to-VideoAlibaba Wan 3.0 Prime video model with text-to-video, image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds with optional synchronized audio.

Wan 3.0 Text-to-VideoAlibaba Tongyi Lab's Wan 3.0 multimodal video model with separate text-to-video, first/last-frame image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds.

Wan 3.0 Image-to-VideoAlibaba Tongyi Lab's Wan 3.0 multimodal video model with separate text-to-video, first/last-frame image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds.

Wan 3.0 Reference-to-VideoAlibaba Tongyi Lab's Wan 3.0 multimodal video model with separate text-to-video, first/last-frame image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds.

FLUX 3 Text to VideoBlack Forest Labs FLUX 3 video model with text-to-video, image-to-video, first-last-frame, extend-video, and keyframes workflows at 720p or 1080p for 5 to 20 seconds, with optional native audio.

FLUX 3 Image to VideoBlack Forest Labs FLUX 3 video model with text-to-video, image-to-video, first-last-frame, extend-video, and keyframes workflows at 720p or 1080p for 5 to 20 seconds, with optional native audio.

MiniMax H3 Max Image to Video API FAQ

What is the MiniMax H3 Max Image to Video API?

MiniMax H3 Max Image to Video is an image-to-video model from MiniMax, optimized through extensive post-training. Using a first-frame image or start and end frames, it follows prompts to generate cinematic 5–15 second videos with native stereo audio and lip sync. Its multimodal spatiotemporal diffusion foundation preserves the subject's identity, clothing textures, and composition while adding coherent, natural motion, lighting, and camera work. Access it programmatically through the API or try it directly in the playground above.

Does MiniMax H3 Max Image to Video support end-frame control?

Yes. Add a second public image URL to the image_urls array to use it as the end frame. The model plans a natural motion path from the opening image to the final composition, making it useful for returning to a target pose, ending on a specific action, or creating smooth transitions.

Can MiniMax H3 Max Image to Video preserve facial identity?

Yes. Post-training closely binds facial features, skin textures, hairstyles, hair strands, and clothing details. As long as the requested movements stay within a natural physical range, the character can closely retain the first frame's identity throughout the shot.

How does MiniMax H3 Max Image to Video generate dialogue and lip movements from a still image?

When the first image shows a clear front or three-quarter view of a face, specify the character's words in the prompt, such as a short English line. In a single generation, the model creates natural lip movements and subtle facial expressions alongside synchronized native speech and ambient sound.

How are resolution and aspect ratio determined in MiniMax H3 Max Image to Video?

The generated video follows the first image's original aspect ratio by default, so no separate aspect_ratio parameter is needed. Choose 480p, 768p, or 1080p; the model scales the image proportionally and adds super-resolution detail within the selected tier.

How is MiniMax H3 Max Image to Video billed?

Billing depends entirely on the requested duration parameter and selected resolution: 480p costs 10 credits per second ($0.050/second), 768p costs 16 credits per second ($0.080/second), and 1080p costs 32 credits per second ($0.160/second). Adding an end-frame image incurs no extra image-processing fee.

What is the difference between Image to Video and Reference to Video?

Image to Video uses the first image as the opening frame and an optional second image as the ending frame, controlling the shot's timeline endpoints. Reference to Video uses images, video, and audio as a feature library, drawing on character appearance, motion rhythm, or style across shots.