MiniMax H3 · Clay Render to Final Render V2V

ABOUT THE WORKFLOW

Render your clay pass

Load an untextured grey clay animation out of Blender and describe the material, lighting and sound you want. The model rebuilds every frame in that style while holding your layout, camera move and timing exactly as animated. You get the finished clip with sound, plus a side-by-side against the original clay pass.

Model

• MiniMax H3 by MiniMax, also called Hailuo 3.0. Open weights under the MiniMax H3 Community License. It reads text, images, video and audio as one input, holds motion from a reference clip frame for frame, and writes stereo sound in the same pass as the picture.

HOW IT WORKS

Step 1. Load your clay video
Your grey untextured render out of Blender, at 24 frames per second. This sets composition, subject positions, camera path and timing.
Works great with: product turntables · character animation · camera fly-throughs

Step 2. Write the look
Describe material, colour, lighting and sound, and keep the Video 1 tag that ties the description to your clay pass. Example: "photorealistic turntable of a modified Honda NCX, azure blue paint, neon underglow, xenon lights, concrete background, photorealistic reflections."

Step 3. Add look plates (optional)
Three image slots sit ready and switched off. Switch one on and add a Picture 1 tag in the prompt when you want a still to set the material and colour rather than words alone.

Step 4. Add audio (optional)
Switched off by default. Pad the recording with silence out to the full clip length so the sound lines up with the picture.

Step 5. Hit run and download
Two files come back: the rendered clip with sound, and a comparison video with the clay pass stitched beside the render.
Ready for: Premiere · After Effects · DaVinci Resolve · Blender

First time? Leave every setting as-is. The defaults (4 steps · 107 frames · 1920x1088) are the right starting point for almost everyone.

RECOMMENDED SETTINGS

Quick-start guide. Find the goal that matches yours and copy the settings.

• Standard render (most people) — 4 steps · 107 frames · seed 11111 · 1920x1088. The right starting point for almost everyone.

• Want a shorter or longer clip — Change how many frames load from the clay video. Length snaps down to a grid: 90, 107, 124, then steps of 17 up to 362. Fifteen seconds is the ceiling in one run.

• Want to reproduce a take you liked — Keep the seed fixed. The same seed with the same prompt returns the same base, so a prompt tweak changes one thing at a time.

• Want a different interpretation of the same shot — Change the seed and leave everything else alone. The layout and camera stay locked to the clay pass either way.

• Want the material to match a specific reference — Switch on a look plate and add a Picture 1 tag in the prompt. Words alone are vague about paint finish, fabric and skin.

• The look is off — Rewrite the prompt before touching a setting. Write instructions, not bans. "One continuous take" lands. "No camera move" does not.

• The timing drifts against your clay pass — Check the clay video is 24 frames per second. Anything else and the motion slides out of sync.

Prompt: Describe the look in one block, then a short timeline, then the audio. Say what to keep and what to change. "Photorealistic turntable, azure blue paint, neon underglow, concrete floor, keep the geometry and animation of Video 1 unchanged" beats "make it look real."

LEARN

📹 Videos

• Intro to Floyo

• ComfyUI 101 Free Course ft. Sebastian Kamph

• Floyo 101 for Team Collaboration

✨ Quick links

• Help docs

• LoRA Training guide

• Discover more workflows

USE CASES

🚗 Product and Vehicle Turntables
Animate a grey model once, then render it in any paint, finish or environment by rewriting the prompt.

🎬 Previs to Final Look
Take an animatic or blocked-out camera move and see it as a finished shot before committing to a full render pipeline.

🏛️ Architectural Walkthroughs
Push an untextured massing model through a camera move and get materials, lighting and ambience without a lighting setup.

🎭 Character Animation Lookdev
Keep your animation timing exactly as you keyed it and test costume, skin and lighting choices in minutes.

🔊 Shots That Need Sound
Describe the audio alongside the visuals and get ambience and effects written in the same pass as the picture.

WHAT WORKS BEST / WHAT TO AVOID

✅ Works great

• Clean grey clay renders at 24 frames per second

• Clear geometry and readable silhouettes

• Prompts that name material, colour, lighting and audio

• Clips at or under 15 seconds

⚠️ May produce softer results

• Clay passes at frame rates other than 24

• Very dark or heavily occluded shots where geometry reads poorly

• Prompts written as bans rather than instructions

• Clips pushed past the 15 second ceiling in one run

FAQ

What is MiniMax H3?
MiniMax H3 is an open-weight video model from MiniMax, the lab behind the Hailuo models, also called Hailuo 3.0. It reads text, images, video and audio as one combined input and generates 4 to 15 seconds of video at 24 frames per second with stereo sound written in the same pass as the picture. This workflow uses the reference variant, which accepts up to 9 images, 3 videos and 3 audio clips.

How does a clay render to final render workflow work?
Your grey untextured clip is passed to the model as a motion and layout reference, not as a starting frame. The model rebuilds each frame with the material, lighting and colour you describe in the prompt, while matching the composition, subject positions, camera path and timing of the clay pass. You get a rendered clip that moves exactly like your Blender animation.

Can MiniMax H3 generate audio with the video?
Yes. Sound is generated in the same pass as the picture, so ambience and effects line up with what happens on screen. Describe the audio in your prompt. You can also upload your own audio clip, padded with silence out to the full clip length.

How long can the video be?
Up to 15 seconds in one run at 24 frames per second. Length snaps down to a fixed grid of 90, 107 and 124 frames, then steps of 17 up to 362. Anything longer needs a separate extend step.

Can I use MiniMax H3 for commercial projects?
The weights are released under the MiniMax H3 Community License rather than Apache or MIT, so commercial use is governed by that license. Some territories are excluded and the 2K pass stays on the API. Review the current license terms before shipping client work.

What is the difference between this and image to video?
Image to video starts from a single still and invents the motion. This workflow starts from a video you animated yourself, so the camera move, staging and timing are yours and the model only supplies the look. If you need control over what happens on screen, animate it in grey first.

How to run MiniMax H3 online?
You can run MiniMax H3 online through Floyo. No installation, no setup, no API key to wire up. Open the workflow in your browser, upload your inputs, and hit run. Free to try.

WHY FLOYO?

Floyo is the only platform with team collaboration for ComfyUI in the browser. You run workflows with no install. You share run history, assets, and models across your team. You pay only when you generate. Floyo supports open-source and closed-source models.

A designer runs an edit and likes the result. A teammate opens that exact run from shared history and keeps going. No file handoffs. No version confusion.

For studios and enterprise teams, Floyo adds private workspaces, pooled resources, and a team usage dashboard. Other ComfyUI cloud tools run for one person at a time. Floyo runs for the whole team, with transparent per-generation costs.

Ready to try it?
Load your clay render, describe the look you want, and run it.

→ Launch Workflow, Free

Questions? Watch the free course or check the FAQ above.