curl-X POST 'https://pollo.ai/api/platform/v1/generation/minimax/minimax-h3/video'\-H'Content-Type: application/json'\-H'x-api-key: YOUR_API_KEY'\-d'{ "input": { "prompt": "A cinematic shot of a golden retriever running through a field of sunflowers at sunset, warm rim light, shallow depth of field", "resolution": "480p", "duration": 5, "aspectRatio": "auto" }}'
선택적으로, 작업 성공 또는 실패 시 콜백을 수신하기 위해 webhookUrl을 포함할 수 있습니다.
선택적 요청 필드
{"input":{"prompt":"A cinematic shot of a golden retriever running through a field of sunflowers at sunset, warm rim light, shallow depth of field","resolution":"480p","duration":5,"aspectRatio":"auto"},"webhookUrl":"https://example.com/webhooks/pollo"}
개요
선택한 엔드포인트 및 모드에 대한 입력 필드입니다.
필드
유형
필수의
화면 설명
prompt
string
예
Text prompt describing the content to generate.
resolution
string
아니요
Output resolution of the generated media (e.g. 720p, 1080p, 2K, 4K). Values are case-insensitive.
MiniMax H3 is a MiniMax video generation model for multi-shot sequences, synced audio, camera control, keyframe editing, and consistent narrative structure.
On Pollo API, developers can access MiniMax H3 for text-to-video, image-to-video, and reference-to-video workflows through a standard API integration.
Key Features of MiniMax H3 API
Multi-Shot Generation
MiniMax H3 supports video outputs composed of multiple connected shots, which can help represent scene progression, actions, or short narrative sequences.
Synced Audio
The model supports native synced audio, allowing prompts to include sound cues, ambience, speech direction, or timing-related audio requirements.
Camera Control
MiniMax H3 allows camera movement and framing to be described in prompts, including shot transitions, angle changes, and subject tracking.
Keyframe Editing
Keyframe editing support helps define important visual states in a video, such as opening composition, ending frame, or major scene transitions.
Asset Referencing
Reference-based generation can use uploaded assets to guide subject appearance, visual style, composition, or continuity across generated clips.
Use Cases of MiniMax H3 API
Narrative Video Clips: Generate short videos with multiple connected scene beats or visual transitions.
Image-to-Video Workflows: Animate still images while keeping the source subject or composition as a reference.
Reference-to-Video Creation: Use reference assets to guide appearance, framing, style, or scene structure.
Product Demonstrations: Create short product clips that show movement, interaction, or step-based usage.
Storyboard Previews: Convert written scene descriptions into moving drafts for planning or review.
Social Video Assets: Generate short clips with camera movement, audio cues, and defined ending frames.
Brand Motion Tests: Produce draft video concepts using reference images, logos, packaging, or visual guidelines.
Creative Tool Integration: Add MiniMax H3 to applications that need prompt, image, and reference-based video generation.
How to Use MiniMax H3 API
Get an API Key: Create a Pollo API account and generate your API key from the developer dashboard.
Choose the Model: Use the MiniMax H3 route, /generation/minimax/minimax-h3, for the generation request.
Add Your Inputs: Provide a prompt, with optional fields such as image, imageTail, resolution, length, and aspectRatio.
Generate and Retrieve: Submit the request, then use the returned taskId or webhook result to check and retrieve the generated video.
Prompting Best Practices for MiniMax H3 API
For MiniMax H3, prompts should define the scene sequence, reference purpose, camera behavior, audio direction, and target ending frame in clear terms.
A Simple Prompt Formula
scene sequence + subject action + reference purpose + camera movement + audio direction + ending frame
What You Should Notice
Scene Order: Describe how the video should begin, progress, and end.
Reference Usage: State whether the reference controls subject identity, composition, style, or final-frame direction.
Camera Direction: Include framing, movement, angle, or transition details when they affect the result.
Audio Requirements: Describe ambience, sound effects, dialogue tone, or timing when synced audio is needed.
Ending Frame: Define the final visual state to make the output easier to evaluate or edit.
Example Prompts
Museum Exhibit Sequence
"Create a short museum exhibit video. Start with a close-up of an artifact case, move to visitors observing the display, add quiet room ambience, and end on a wide shot of the exhibition title wall."
Cooking Clip From Image
"Use the uploaded dish image as the final plating reference. Show vegetables being prepared, sauce added to a pan, steam rising, and an overhead camera movement ending on the completed dish."
Game Environment Preview
"Generate a short game environment preview. Begin at a closed stone gate, track forward as light appears through the cracks, cut to an underground hall, add low ambience, and end on a glowing checkpoint."
Character Reference Scene
"Use the reference image for character appearance. Show the character entering a workshop, turning on a desk lamp, picking up a small mechanical part, and ending with a close-up under warm light."
Product Assembly Clip
"Create a compact product assembly video. Show a foldable lamp base opening, the arm extending, the shade attaching, and the light turning on, with small mechanical sounds and a final side-view frame."
MiniMax H3 vs Hailuo 2.3 vs Google Veo 3.1
Capability
MiniMax H3
Hailuo 2.3
Google Veo 3.1
Core Focus
Multi-shot video generation with synced audio
General video generation in the Hailuo model family
Video generation with strong audiovisual control
Input Workflows
Text, image, and reference-to-video
Text and image-to-video
Text and reference-guided video
Audio Support
Native synced audio
Visual-first generation workflows
Audio-video generation support
Camera Control
Prompt-based camera and framing control
General motion prompting
Detailed camera and scene direction
Sequence Control
Supports connected narrative sequences
Suited to shorter direct clips
Suited to polished scene generation
Recommended For
Multi-shot clips, reference workflows, and audio-aware video
Fast visual video creation
Higher-end audiovisual scene generation
Why Choose MiniMax H3 API?
MiniMax H3 is suitable for workflows that require multi-shot structure, audio timing, camera direction, and reference-guided video generation.
Through Pollo API, developers can access MiniMax H3 alongside 300+ image and video models, with documentation, pricing, infrastructure, and support in one platform.
Use MiniMax H3 on Pollo API to build reference-aware video tools, ecommerce generators, creator apps, and brand production workflows that need controllable short-form video with sound.
MiniMax H3 API FAQs
What is MiniMax H3?
MiniMax H3 is a general-purpose multimodal video generation model from MiniMax. It understands text, image, video, and audio context to produce short video with native stereo sound.
What input types does MiniMax H3 support?
MiniMax H3 supports text-to-video, image-to-video, and reference-to-video workflows.
What makes MiniMax H3 different from Hailuo 2.3?
MiniMax H3 is designed around unified multimodal context, native stereo output, and broader reference/editing workflows. Hailuo 2.3 is better understood as an earlier visual video model.
Can MiniMax H3 use audio or video references?
Yes. MiniMax H3 can use reference images, videos, and audio to guide character identity, motion transfer, camera behavior, voice, style, or editing rhythm in the generated result.
What is MiniMax H3 suitable for?
MiniMax H3 is suitable for multi-shot videos, image animation, reference-guided clips, storyboard previews, product demos, and social video assets.
Why run MiniMax H3 through Pollo API?
Pollo API gives one integration for MiniMax H3 and 300+ other models, with documentation, task tracking, and lower-cost generation than comparable providers.
