- One image and speech audio
Inputs
- Follows the audio
Length
- 480P to 2K
Resolution
- 768P
Default
About MiniMax H3 Max Lip Sync
MiniMax H3 Max Lip Sync is the talking-head member of the MiniMax H3 Max video family. You supply two things: a still image of the person who should speak and an audio file of what they say. The route animates the face so the mouth follows the soundtrack, and the clip runs as long as the audio does, so there is no duration setting to choose.
It sits next to the other H3 Max routes in Masonry for a different job. MiniMax H3 Max generates a scene from text, a frame, or reference images and writes its own sound. Lip Sync starts from sound you already have, such as an approved voiceover, a translated script, or a recording from your own spokesperson, and only has to make a face deliver it.
fal, which serves the route, says the output uses the supported aspect ratio nearest your image and that the image's aspect ratio must sit between 0.4 and 2.5, so both a vertical phone portrait and a wide frame work. Transcription is on by default: the audio is transcribed to guide the mouth shapes. fal advises switching it off for audio with a lot of non-word sound, such as music, singing, or laughter, because the transcript can introduce words that were never spoken. This page describes the route from Masonry's catalog and fal's documentation; we have not published a first-hand test of it yet.
Why teams choose MiniMax H3 Max Lip Sync
Choose MiniMax H3 Max Lip Sync when the words are already decided and recorded, and the job is to give them a face. Because the audio drives both the mouth and the length, the clip matches your approved script rather than one the model wrote. Use a portrait of someone who has agreed to appear and to have their likeness animated, and label synthetic video where your channel or law requires it. If you need the model to write the scene and the sound together, use MiniMax H3 Max instead.
What MiniMax H3 Max Lip Sync can do
The capabilities that set MiniMax H3 Max Lip Sync apart and earn its place in a brief
Your Audio Sets the Length
There is no duration setting. The clip runs as long as the speech file you supply, so a 20-second line becomes a 20-second video.
One Portrait In
A single still image of the speaker is the only visual input. Use a clear, front-facing image of the person who has agreed to appear.
480P to 2K Output
Choose 480P, 768P, 1080P, or 2K. 768P is the default.
Transcription-Guided Mouth Shapes
The speech is transcribed to guide lip movement. Turn this off for music, singing, laughter, or noisy audio, where a transcript can invent words.
Wide Range of Image Shapes
Per fal, the image's aspect ratio can be anywhere from 0.4 to 2.5, and the output uses the nearest supported ratio.
Pairs With the H3 Max Family
Generate a scene with MiniMax H3 Max, then use Lip Sync for the shots where a specific person must deliver a specific recorded line.
Where teams reach for MiniMax H3 Max Lip Sync
- Spokesperson clips from an approved voiceover and a photo of that spokesperson
- Translated versions of the same message, one audio file per language
- Product explainers where the script is fixed and recorded before the visuals
- Founder or support-team updates recorded as audio first
- Training and onboarding clips with a consistent presenter
- Short social clips built from a podcast or interview excerpt
Frequently asked questions
What teams need to know about creating with MiniMax H3 Max Lip Sync in Masonry
What is MiniMax H3 Max Lip Sync?
It is a route in the MiniMax H3 Max video family that animates a single portrait so it speaks along to an audio file you supply. In Masonry it is listed under the key minimax-h3-max-lipsync and takes one image and one speech audio file.
How long can a Lip Sync video be?
The clip follows the length of your audio. fal documents a minimum of 5 seconds of audio and support for long recordings, so split very long scripts into scenes you can review one at a time.
What resolutions does it support?
480P, 768P, 1080P, and 2K, with 768P as the default. The output uses the supported aspect ratio closest to your image.
Should I turn transcription off?
Leave it on for clean speech, where the transcript helps the mouth shapes follow the words. fal advises turning it off for audio with a lot of music, singing, laughter, or noise, because the transcript can introduce words that were never said.
Can I use a photo of any person?
Only use a portrait of someone who has agreed to appear and to have their likeness animated with their voice or an approved voice. A talking clip of a real person who did not consent can mislead viewers and can break platform rules and the law where you publish.
How is it different from MiniMax H3 Max?
MiniMax H3 Max generates a scene, its motion, and its sound from text, a frame, or reference images. Lip Sync starts from audio you already have and a single portrait, and only animates the face to deliver that audio.
Has Masonry tested this route?
Not yet. This page is written from Masonry's live catalog and fal's documentation, checked November 18, 2026. Run a short clip on your own portrait and audio before relying on it for a campaign.
How does my team use MiniMax H3 Max Lip Sync in Masonry?
Open a Masonry canvas, pick MiniMax H3 Max Lip Sync from the model selector, and describe the video you need: a product shot, an ad creative, a social post. Masonry generates it, then you refine, edit, and combine MiniMax H3 Max Lip Sync with other models in one workspace.
Is MiniMax H3 Max Lip Sync free to try?
New Masonry accounts get 50 free credits to try GPT Image 2.5 Flare. MiniMax H3 Max Lip Sync is included in every paid plan, and the editor shows the exact credit cost before you generate.
How do I write good prompts for MiniMax H3 Max Lip Sync?
There is no text prompt on this route; the image and the audio do the work. Pick a sharp, evenly lit portrait with the mouth clearly visible and nothing covering it. Record clean speech with little background noise, and switch transcription off when the audio includes music, singing, or laughter. Check the finished clip for mouth drift on fast or overlapping words before you publish.
Who makes MiniMax H3 Max Lip Sync?
MiniMax H3 Max Lip Sync is built by fal-ai. Inside Masonry it runs alongside 50+ image and video models, so your team can pick the right one for each brief without switching tools.
Start creating with MiniMax H3 Max Lip Sync
Generate, edit, and compare across 50+ models in one workspace.
Guides for MiniMax H3 Max Lip Sync
Prompt walkthroughs and examples from the Masonry blog
Explore more AI models
Compare MiniMax H3 Max Lip Sync with other models teams run in Masonry