MiniMax H3, also known as Hailuo 03, unifies text, image, video, and audio for 2K video generation, multimodal referencing, native stereo sound, motion transfer, and precise editing.
Video description prompt.
0/7000Video duration in seconds. Available values: 4-15.
The ratio of output video.
Video resolution.
Explore different use cases and parameter configurations
Click to upload or drag and drop images.
Images: JPEG/JPG/PNG/WebP, Max 30MB
Video description prompt
0/7000Video duration in seconds. Available values: 4-15.
Video resolution.
Explore different use cases and parameter configurations
Video description prompt.
0/7000Click to upload or drag and drop
Supported formats: JPEG, PNG, WEBP, HEIC Maximum file size: 30MB; Maximum files: 9
reference images.
Click to upload or drag and drop
Supported formats: MP4, QUICKTIME Maximum file size: 50MB; Maximum files: 3
reference videos.
Click to upload or drag and drop
Supported formats: MPEG, WAV, X-WAV Maximum file size: 15MB; Maximum files: 3
reference audios.
Video duration in seconds. Available values: 4-15.
The ratio of output video.
Video resolution.
Explore different use cases and parameter configurations
Complete guide to using MiniMax H3 (Hailuo-03)
Use MiniMax H3 API on Kie.ai to turn prompts and multimodal references into polished 2K videos with native stereo sound, consistent motion, and precise creative control.
MiniMax H3, also known as Hailuo 03, is an open, general-purpose multimodal video generation system that understands text, images, video, and audio within a unified creative context. It produces 5- to 15-second videos at up to 2K resolution with native stereo audio, while following detailed instructions for characters, actions, camera movement, shot structure, visual style, dialogue, music, sound effects, readable text, and brand elements. Unified multimodal understanding also enables reference-based creation, audiovisual transfer, and precise editing for commercial content production.MiniMax H3 API is now available on Kie.ai through three generation methods: Text to Video, Image to Video, and Reference to Video. Developers can select a suitable mode based on available source materials, whether starting from a written prompt, animating a single image or defined first and last frames, or combining multiple image, video, and audio references. View latest MiniMax H3 API documentation on Kie.ai for current request parameters, supported input formats, generation settings, and integration requirements.
Text to Video creates a complete audiovisual clip directly from a written prompt without requiring uploaded media. Developers can define the environment, characters, actions, camera direction, shot progression, visual treatment, dialogue, music, and ambient sound in one request. MiniMax H3 API suits advertising concepts, cinematic sequences, animated shorts, motion graphics, social content, and product visuals created entirely from textual direction.
Image to Video supports either one uploaded image or a pair of first and last frames. A single image can act as the starting visual state for natural character motion, environmental movement, and camera animation. Providing both opening and closing frames gives developers clearer control over how the sequence begins, transitions, and concludes. Hailuo 03 API works well for animated posters, product showcases, UI demonstrations, character performances, key-art animation, and structured visual transitions.
Reference to Video combines written instructions with reference images, video clips, and audio files in one generation request. Developers can assign individual assets to character identity, product appearance, motion, camera language, editing rhythm, visual style, voice, or sound design. Hailuo-03 API also supports subject replacement, motion transfer, background modification, relighting, dialogue changes, voice matching, wardrobe updates, product swaps, and localized edits while preserving unaffected parts of the original footage.
MiniMax H3 API can interpret text, images, video clips, and audio recordings within the same context. Each input can serve a different role, such as defining character identity, guiding movement, establishing camera language, matching a visual style, or providing voice and sound references. This unified understanding helps different creative elements remain connected throughout the final video.
Hailuo 03 API generates up to 2K video together with native stereo sound, rather than adding audio through a separate process. Visual action, dialogue, music, ambience, and sound effects can be directed within the same prompt, producing stronger audiovisual continuity and more natural synchronization between what appears on screen and what is heard.
Hailuo-03 API can combine multiple images, video clips, and audio files within one generation. Identity may come from an image, motion and camera behavior from a video, and voice or sound characteristics from an audio reference. Natural-language instructions clarify how every asset should influence the result, reducing conflicts between different reference materials.
Hailuo 3.0 API follows detailed directions for shot order, timing, transitions, camera movement, character behavior, sound design, and elements that must remain unchanged. It also renders clearer text, titles, interface components, signage, and brand details, while responding more reliably to both positive instructions and explicit restrictions.
Hailuo AI API can apply targeted changes to existing footage without unnecessarily rebuilding the entire scene. Characters, objects, clothing, backgrounds, lighting, dialogue, voices, and visual effects can be replaced or adjusted while preserving unaffected composition, movement, timing, and environmental details.
Given similar creative direction, MiniMax H3 and Seedance 2.0 often arrive at noticeably different visual solutions. The four comparisons below highlight their approaches to product storytelling, motion design, cinematic visualization, and interface animation.
MiniMax H3 builds a more structured commercial through varied camera angles, close-up feature shots, connected charging actions, readable specifications, and consistent product materials across the sequence. Seedance 2.0 creates a polished studio presentation with smooth movement and clean lighting, but follows a simpler shot progression with less emphasis on individual functions. MiniMax H3 maintains stronger continuity between product design, feature communication, typography, and the final brand reveal.
Hailuo 03 develops a denser motion-design sequence through grids, fluid forms, dimensional objects, bold typography, and rapidly changing compositions. Seedance 2.0 favors cleaner geometric transformations, more restrained layouts, and a simpler progression toward the closing title card. Hailuo 03 brings greater visual variety and editorial energy, while Seedance 2.0 maintains a more minimal and uniform graphic language.
Hailuo-03 presents the racing car through realistic materials, controlled close-ups, technical overlays, and a consistent black, white, and red visual system across changing viewpoints. Seedance 2.0 leans further into glowing blueprint effects, holographic components, and a stronger impression of speed. Hailuo-03 keeps the vehicle identity clearer and more commercially grounded, while Seedance 2.0 places greater emphasis on stylized technical spectacle.
Hailuo 3.0 keeps the controller integrated with the original webpage, preserving its typography, carbon-fiber textures, interface controls, and high-contrast visual hierarchy as the camera moves through the layout. Seedance 2.0 transforms the material into a standalone product presentation with blue motion backgrounds, dramatic close-ups, and newly introduced promotional text. Hailuo 3.0 remains closer to the original interface structure, whereas Seedance 2.0 takes a freer approach that resembles a conventional product advertisement.
Kie.ai makes it easier to move from initial experimentation to a complete video generation experience. Test creative settings first, define a reliable input structure, and then bring the validated workflow into your product.
Product films can combine close-up details, feature demonstrations, branded typography, camera transitions, and synchronized sound within one continuous sequence. MiniMax H3 API also helps preserve product materials, proportions, packaging, and visual identity across changing shots, making the final result feel more cohesive and commercially polished.
Hailuo 03 API can build opening sequences, animated posters, music visuals, and editorial motion graphics from detailed creative direction. Typography, geometric layouts, layered effects, transitions, camera movement, music, and sound cues can develop together over time instead of appearing as disconnected visual elements.
Short dramas, character promos, motion comics, and stylized narratives benefit from consistent faces, clothing, props, voices, and performance details across the scene. Hailuo-03 API can also draw motion, camera language, atmosphere, or dialogue delivery from separate references while keeping the central character recognizable.
Interface animations can preserve webpage structure, menus, typography, controls, and product imagery while introducing scrolling, hover states, transitions, loading sequences, and environmental reveals. Hailuo 3.0 API also supports game menus, equipment customization, cinematic gameplay concepts, and other digital experiences that combine interface logic with animated storytelling.
Strong results depend on more than describing what should appear on screen. Clear prompts should organize visual intent, reference roles, timing, camera direction, sound, and preservation requirements so every part of the generation follows the same creative plan.
When several images, videos, or audio files are used together, explain what each asset should control. Separate references can define character identity, product appearance, motion, camera language, editing rhythm, visual style, voice, or sound, reducing conflicts between different inputs.
Break multi-shot ideas into a clear progression so actions, transitions, and visual changes happen in the intended order. A structured sequence helps Hailuo 03 API maintain pacing across the full duration instead of compressing every idea into the opening seconds or producing disconnected scenes.
Treat dialogue, music, ambience, and sound effects as part of the same creative direction as composition and movement. Hailuo-03 API can coordinate audio with visible actions more effectively when sound cues, emotional tone, and changes over time are described alongside the visual sequence.
List the facial features, clothing, props, materials, colors, typography, or product details that must remain consistent. Clear preservation instructions give Hailuo 3.0 API stronger anchors for maintaining recognizable subjects and visual elements throughout changing shots and camera angles.
Describe framing, lens behavior, camera movement, focus changes, exposure, and overall photographic treatment. Specific cinematography direction helps MiniMax H3 API produce more intentional visual storytelling and prevents the sequence from relying on generic movement or repetitive compositions.
State which elements, transitions, styles, or visual errors should be avoided. Negative constraints help Hailuo 03 API stay within the intended creative direction, especially when working with typography, character identity, interface layouts, realistic footage, or highly stylized scenes.
For video editing, identify each required change and clarify what must remain untouched. Separating replacements, removals, relighting, dialogue changes, and background adjustments allows Hailuo AI API to apply localized modifications while preserving the original timing, composition, movement, and unaffected scene details.
Flexible usage-based pricing makes it easier to control costs across testing, recurring content production, and higher-volume generation. MiniMax H3 API pricing on Kie.ai supports gradual scaling without requiring dedicated hardware or complex infrastructure investment.
Detailed documentation covers authentication, generation modes, supported inputs, request settings, and implementation guidance. Hailuo 03 API resources help reduce unnecessary setup work and make each stage of integration easier to understand.
Playground provides a visual environment for testing Text to Video, Image to Video, and Reference to Video before beginning full integration. Prompts, single images, first and last frames, and multimodal references can be adjusted and compared more efficiently.
Stable service access supports consistent generation across development, testing, deployment, and ongoing production. Hailuo 3.0 API infrastructure on Kie.ai is built to handle growing request volumes without requiring users to maintain their own video-generation environment.
Professional support is available to assist with account access, documentation questions, integration issues, and generation-related concerns. MiniMax H3 API users can receive timely guidance throughout testing and production instead of resolving technical problems alone.
Kie.ai provides access to a broader collection of video, image, audio, and language models through one account. Beyond Hailuo AI API, users can compare different creative capabilities, expand existing products, and add new generation workflows without rebuilding the entire technical foundation.