AI Cut and Big Speed Improvements
A new feature that edits your talking-head videos for you. Audiobooks, dubbing, and clips are also a lot faster.
Here's what's new in Voice Creator Pro v2.3.9. Everything below is live on both the desktop app and in VCP Cloud.
New Feature: AI Cut
AI Cut is a new tool for editing videos where you're talking to the camera. Give it a raw recording and it removes long silences, filler words like um and uh, stutters, false starts, and bad takes. If you say a line more than once, it keeps the last good take and cuts the earlier ones. If you restart a whole section, it keeps the final run-through.
You can also give AI Cut your script and it will keep the take closest to it. There's a strict mode that also cuts anything you said that isn't in the script, and any script lines it couldn't find in the video get flagged.
You can review every cut and see why each take was cut, restore anything you want back, and make your own cuts.
- Automatic cuts for silences, filler words, and bad takes, keeping your best take of every line
- Subtitles with the same styles as Dubbing and Viral Clips, burned into the export
- Word editing to fix a misheard word, highlight it, or animate it
- Any aspect ratio (original, 9:16, 1:1, 4:5, or 16:9) with the crop following your face
- Premiere Pro and DaVinci Resolve XML and SRT subtitle exports
Free plans get one AI Cut video a month, up to 5 minutes, with a watermark. On desktop you bring your own AI model key (Gemini, OpenAI, OpenRouter, DeepSeek, or any OpenAI-compatible endpoint). In the cloud it's included.
Performance Improvements
We spent a lot of this release making things faster. On desktop that means less time waiting on your own computer. In the cloud you get more usage for your money.
- Audiobooks generate up to 3x faster with Qwen3 and more than 2x faster with Kokoro on NVIDIA cards
- Dubbing is about 1.5x faster, and faster still with Qwen3 voices
- Lip sync is up to 2x faster
- Viral Clips render about 3x faster
- Faster video exports on AMD and Intel graphics cards for Viral Clips and dubbing
- DramaBox is faster on 12 GB cards and uses much less system memory
Other Improvements
- Viral Clips editing lets you remove words, trim and cut clips, cut filler words, pick an aspect ratio, and use new layouts for three or four speakers or a screen share
- Nemotron speaker detection for Speech to Text, Dubbing, and Clips, better at quick back-and-forth
- Crosstalk in dubs is kept when two people talk at once, and dubbed lines can overlap like the original
- Audiobook MP3s now include the title, author, narrator, cover, and chapters
- Podcast loudness option for audiobook exports
- One TTS tab in the Lab, with Clone and TTS merged
- Orukeet, a more accurate transcription model across accents and languages
- Subtitles can export as JSON, and get italic, underline, and the Alumni Sans font
- Intel Arc now runs transcription, dubbing, and clips
- Many smaller fixes and improvements