MCP / Models / Reference video, editing and lip sync
MiniMax H3 Max Reference to Video MCP: up to nine references
MiniMax H3 Max makes a 1 to 15 second video that keeps the characters, products and places of up to nine reference pictures. Run it from Claude Code, Codex, Cursor or any MCP client through one FlyMy.AI MCP, or call its API.

$ YOUR NEXT RUN
Use MiniMax H3 Max Reference to Video in your assistant
Paste this prompt into Claude Code, Codex, Cursor or any agent. It reads the guide and helps you connect.
Read https://flymy.ai/skill.md and https://flymy.ai/mcp/models/minimax-h3-max-reference-to-video.md. Help me connect the FlyMy.AI MCP to this assistant, then run MiniMax H3 Max Reference to Video (flymyai/minimax-h3-max-reference-to-video) for this task: Use MiniMax H3 Max Reference to Video to make a clip of the character in my reference images walking along a desert road at dusk.
What MiniMax H3 Max Reference to Video does
Send prompt and images, both required: one to nine reference pictures (characters, products, places or a style), each with an aspect ratio between 2:5 and 5:2, named Image 1, Image 2 and so on in the prompt in the order you send them. image_url fixes the first frame and end_image_url the last. duration is 1 to 15 seconds (5 by default), resolution is 480P, 768P (the default) or 1080P, aspect_ratio is adaptive, the default, or 21:9, 16:9, 4:3, 1:1, 3:4 or 9:16, and prompt_expansion_mode and seed work as in the other H3 Max models. Outputs: video, one MP4 clip, expanded_prompt and reference_tokens, the reference tokens the run counted.
It is billed per second of video you ask for, more at a higher resolution, plus the reference pictures beyond a free allowance; the price shown before the run is an upper bound held on your balance, and you are charged what the run cost. It takes pictures only, no reference videos, reference audio, middle keyframe or soundtrack of your own, and does not make a video from text alone (MiniMax H3 Max Text to Video does).
If a picture shows a real person, use it only with their consent.
MiniMax H3 Max Reference to Video inputs
Every input of flymyai/minimax-h3-max-reference-to-video as its live schema lists them. Returns video (array), expanded_prompt (string), reference_tokens (string).
| Input | Type, values and default |
|---|---|
# prompt | string required Describe the video and name the pictures in it as Image 1, Image 2, ... in the order you send them |
# images | string required One to nine reference pictures (subjects, characters, products, a style), links or uploads, named Image 1, Image 2, ... in the prompt; each one's aspect ratio must be between 2:5 and 5:2. The first 4,096 reference tokens are free (a square picture counts 1,024, 4:3 1,376, 16:9 1,824, up to 2,560); beyond them 1,000 tokens cost $0.02. Use a real person's picture only with their permission |
# image_url | file optional Allowed: image file, as a URL or upload Optional first frame: the video opens exactly on this picture, a link or an upload, and takes its shape; it counts about 600 to 1,400 reference tokens ($0.02 per 1,000 beyond the free 4,096) |
# end_image_url | file optional Allowed: image file, as a URL or upload Optional last frame: the video ends exactly on this picture, a link or an upload; it counts about 600 to 1,400 reference tokens ($0.02 per 1,000 beyond the free 4,096) |
# duration | integer optional Allowed: 1 to 15 Default: 5 Length of the clip in seconds, 1-15; the price is per second ($0.05 at 480P, $0.08 at 768P, $0.16 at 1080P). The finished clip can run up to about 0.7 seconds longer than you ask for; the price counts the seconds you ask for |
# resolution | string optional Allowed: 480P, 768P, 1080P Default: 768P 480P ($0.05 a second), 768P ($0.08) or 1080P ($0.16, refined from a native 768P render) |
# aspect_ratio | string optional Allowed: 16:9, 4:3, 1:1, 3:4, 9:16, 21:9, adaptive Default: adaptive Shape of the video: adaptive (the default: the model picks it, or follows your first frame), 21:9, 16:9, 4:3, 1:1, 3:4 or 9:16 |
# prompt_expansion_mode | string optional Allowed: disabled, balanced, quality Default: balanced How much effort goes into rewriting your prompt before the video is made: disabled (your words as they are), balanced (about a second, the default) or quality (up to about 30 seconds, a richer prompt). No price effect |
# seed | integer optional Allowed: 0 to 2147483647 Seed for the generation; a random one is used when you leave it out |
Ask your agent
Requests to paste as they are, once the FlyMy.AI MCP is connected.
flymyai/minimax-h3-max-reference-to-videoflymyai/minimax-h3-max-reference-to-videoflymyai/minimax-h3-max-reference-to-videoMiniMax H3 Max Reference to Video price
From $0.05 per second of duration, depending on resolution, images, image_url and end_image_url. A run at the default settings costs $0.40. A run costs $0.05 to $2.49, depending on duration and resolution and images and image_url and end_image_url. Billed to your FlyMy.AI balance; no seats, no subscription and no minimum.
| Setting | Price |
|---|---|
| resolution 480P | $0.05 per second |
| resolution 768P | $0.08 per second |
| resolution 1080P | $0.16 per second |
Call MiniMax H3 Max Reference to Video from code
The same model, from an agent or from your own code. Keys stay in your FlyMy.AI account.
From an agent, through the MCP
The agent calls get_model_schema, then run_model with:
{
"endpoint_id": "flymyai/minimax-h3-max-reference-to-video",
"input": {
"prompt": "Keep the subject and its motion, set the scene on a beach at golden hour",
"images": [
"https://example.com/front.jpg",
"https://example.com/back.jpg"
]
}
}From your code, over HTTP
Form fields, not JSON; files go as uploads or public URLs. The answer is a prediction_id; read the result at https://api.flymy.ai/api/v1/flymyai/minimax-h3-max-reference-to-video/predict/async/result/?request_id=<prediction_id>.
curl -X POST https://api.flymy.ai/api/v1/flymyai/minimax-h3-max-reference-to-video/predict/async/ \
-H "X-API-KEY: $FLYMYAI_API_KEY" \
-F 'prompt=Keep the subject and its motion, set the scene on a beach at golden hour' \
-F 'images=["https://example.com/front.jpg", "https://example.com/back.jpg"]'Questions about MiniMax H3 Max Reference to Video
Do I need a MiniMax API key for MiniMax H3 Max Reference to Video?
No. MiniMax H3 Max Reference to Video runs on FlyMy.AI's account, so you need no MiniMax key or plan: sign in to FlyMy.AI once and each run is billed to your FlyMy.AI balance.
How much does MiniMax H3 Max Reference to Video cost?
From $0.05 per second of duration, depending on resolution, images, image_url and end_image_url. There is no subscription or minimum; the price comes from FlyMy.AI billing and was checked on 2026-11-18.
How do I use MiniMax H3 Max Reference to Video in Claude Code, Codex or Cursor?
Add the FlyMy.AI MCP with one command and sign in, then ask in plain words, for example: "Use MiniMax H3 Max Reference to Video to make a clip of the character in my reference images walking along a desert road at dusk." Your agent reads the inputs with get_model_schema and runs flymyai/minimax-h3-max-reference-to-video with run_model.
What does MiniMax H3 Max Reference to Video need as input?
MiniMax H3 Max Reference to Video requires prompt and images; it also takes 7 optional inputs, each listed on this page with its allowed values and default.
Can I call MiniMax H3 Max Reference to Video from my own code?
Yes. Send form fields to https://api.flymy.ai/api/v1/flymyai/minimax-h3-max-reference-to-video/predict/async/ with your FlyMy.AI API key in the X-API-KEY header.
Go to the source
From the FlyMy.AI model catalog of 2026-11-18: inputs from the model's live schema, prices from FlyMy.AI billing. Requests are examples to paste, not recorded runs.





