Prophetero Take Mixer [GLOBAL]

Keep the best picture from one video and the best sound from another. Take Mixer lets you choose separate start times, check the result in a short preview, and export a new MP4.

This is a ComfyUI workflow with a custom node, not a generation model. It works with videos made by LTX, MiniMax H3, or other tools. No checkpoint, LoRA, GPU, or cloud account is required for mixing.

What it does

• Select one video for the picture and another for the sound.

• Adjust their start times independently to align the action and audio.

• Preview up to four seconds before exporting.

• Export an H.264/AAC MP4 at the source size, or fit or crop it to 1920×1080.

• Use the interface and workflow notes in eight languages; English is the default.

Install and use

1. Download and fully extract the ZIP.

2. Stop ComfyUI and run START_HERE.cmd. Select the ComfyUI folder you actually use—including the backend folder if you launch it through SwarmUI.

3. Restart ComfyUI or SwarmUI, then open Prophetero_Take_Mixer_GLOBAL.json.

4. Choose your picture and sound videos, set the start times and duration,
   
   Not just for H3. Prophetero Take Mixer works with videos from H3, LTX, other AI generators, and cameras. Take the picture from one video and the sound from another—no specific model or LoRA is required. It supports local SDR video files.