None
None
Long Story Video Skill
Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.
Ads Video Skill
Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.
3D Science Explainer Video Skill
Convert scientific concepts into stunning 3D explain animations
Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
Minimax H3 Video To Video
Turn a finished take into a different cast: upload a few photos and Minimax H3 Video To Video puts new faces into footage you already shot.
All Tools
Discover our comprehensive AI-powered animation toolkit
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator
Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
What Minimax H3 Video To Video Changes — and What It Leaves Alone
Think of it as a re-casting pass over an edit you are already happy with: the people on screen change, while the performance, pacing and sound you approved stay exactly where they were.
- Not another text-to-video modelIt never invents a scene from scratch. Instead it works over footage that already exists, so the original look, blocking and framing are never rebuilt from zero.
- Up to four new faces, one photo eachSupply a single clear picture per performer — four is the maximum — and those faces claim the main on-screen roles, tracking them from left to right wherever they appear.
- Your footage keeps its performanceGestures, timing, lighting, lens behaviour, edit rhythm and the existing soundtrack all survive the pass, so the swapped version drops back into your timeline without a rebuild.
Recast a Clip in Three Steps
Start with footage you are happy with, add one photo per new performer, then let the model trade out the cast while everything else holds.
Built for Clips That Already Work
Only the faces change. The performance, the rhythm of the edit and the audio mix you signed off on come through the pass exactly as they were.
Four performers, mapped left to right
Four photos is the ceiling, and they step into the main roles in the order they appear on screen. Each identity is followed through cuts, and a short prompt can shuffle the assignment.
Acting is inherited from your footage
Body language, pacing, expressions, lighting and camera motion are inherited from the source clip, while the new faces and clothing are pulled from the pictures you upload.
Edit and audio left untouched
Your cut points and the existing audio track pass through unchanged, which means the finished file drops into a locked sequence instead of needing to be rebuilt.
Low-cost preview, high-res finish
Run a 768p check to confirm the swap before paying for the higher resolution, and since runtime mirrors the source, cost scales with clip length rather than resolution.
Hair and wardrobe come from the photo
If your reference shows more than a face, the hairstyle and clothing travel with it — which is why a half-body frame outperforms a tight headshot.
Made for takes that are already right
Film crews, brand teams, social creators and fashion or commerce studios reach for it whenever a take is flawless but the wrong performer is standing in the frame.
Minimax H3 Video To Video: Common Questions
Answers to the questions that come up before swapping the cast in an existing clip.
What is Minimax H3 Video To Video used for?
It gives an existing video a new cast. The people in your photos step into the roles played in the source, while the original movement, camera behaviour, edit and sound are left alone. By default the photos take the main roles in left-to-right order wherever those roles appear.
Is there a length limit on the source video?
Between 5 and 30 seconds, and no single shot should run past 15 seconds, so cut longer material first. The recast you get back is the same duration as what you fed in.
How many performers can I replace at once?
Four at most, with one photo per person. They claim the leading on-screen roles from left to right unless you say otherwise — a prompt like 'the woman on the right becomes person 1' rewrites that mapping.
Which reference photos give the best results?
One crisp, front-facing, evenly lit image per person. A half-body frame carries hair and clothing better than a tight headshot. Steer clear of sunglasses, hats, hands covering the face, group shots, and low-resolution or heavily filtered portraits.
Do I need 768p or 1080p?
Start at 768p to confirm the swap reads correctly, then render the 1080p master. Cost tracks how long the clip runs more than it tracks resolution.
When is this the wrong tool?
Avoid it when you need new movement, new camera angles or a different setting. It also struggles with crowded scenes, small or distant faces, fast action and occlusion after hard cuts, and identity can drift back toward the original performer in long shots or profile turns.
Start Recasting with Minimax H3 Video To Video
Insert new performers into footage that is already edited, and keep the timing, camera work and audio exactly as they were. Then browse the recast guide or compare the related MiniMax, Hailuo and Kling models.