SpeechSwap – OmniVoice Video Dialogue Replacement - v1.0 Showcase

Use this as the post description:

SpeechSwap v1.0 – ComfyUI Video Dialogue Replacement

SpeechSwap replaces dialogue inside an exact video time range using OmniVoice voice cloning while preserving the original video, stereo ambience, music, and sound effects.

Features

• Synchronized video and waveform editor

• Millisecond-accurate start/end selection

• OmniVoice voice cloning

• Whisper transcription

• UVR dialogue/background separation

• Stereo background preservation and level matching

• Automatic speech-duration fitting

• Preview before committing an edit

• Saved-version history and rollback

• Iterative replacement of multiple dialogue sections

• Optional lossless WAV/FLAC export for external lip-sync

Installation

1. Download and extract SpeechSwap-portable.zip.

2. Read QUICK_START.txt.

3. Close ComfyUI.

4. Run Install-SpeechSwap.bat.

5. Enter your ComfyUI folder.

6. Restart ComfyUI and open SpeechSwap.workflow.json.

The installer downloads and configures the required custom nodes, Python dependencies, UVR environment, and dialogue-separation model.

For large videos, launch ComfyUI with:

--max-upload-size 4096

Important

The ZIP does not include model weights, source videos, reference recordings, or generated media. Internet access is required during installation.

This workflow does not perform lip-sync internally. It can export the completed soundtrack for use with an external lip-sync application.

Only clone voices and edit recordings when you have permission. Users are responsible for respecting third-party licenses and obtaining the necessary rights to source media and voices.

ZIP SHA-256

D1D4EB9E54AF370D2A5FA5CE0BD2E46C331C1B0E88830F471800AB43A5E79B18