Turn Any Long Video Into Short Vertical Clips

Here's what's new in Voice Creator Pro v2.1.8. The headline is a completely new feature, and the two most expensive ones are no longer paid-only. Everything below is live on both the desktop app and in VCP Cloud.

Viral Clips

Picking the best sixty second clips out of a two-hour recording is extremely tedious, so VCP can now make use of large language models to make that effortless. Drop in a podcast, an interview, or a two-hour stream, and get back short vertical clips ready to post.

• It picks the moments. A two-stage pass reads the transcript and finds the most viral moments. You can type in guidance about what you're looking for and narrow the video range before it starts.
• It reframes vertically, and it follows whoever is talking. Face and active-speaker tracking, rule-of-thirds framing, real cuts between speakers, and a dynamic split screen when two people are talking over each other.
• It edits. Punch-ins anchored to important words, colour effects snapped onto scene cuts, and split edits that hold the picture through a pause. Every effect lands on a timeline you can drag, retime, add to, or re-roll.
• Export a folder per clip carrying the video, a Premiere Pro XML timeline already framed vertically, episode and clip SRTs, and the source.

Import from a file or straight from a YouTube URL (desktop only). On desktop you bring your own model key (Gemini, OpenAI, OpenRouter, DeepSeek, or any OpenAI-compatible endpoint); in the cloud it's included.

Dubbing and Clips on the Free Plan

Both used to be unavailable on the free plan. They're metered by the same monthly allowance as everything else now. Free runs come back watermarked, dubbing caps at 720p, and clips cap at one per run, so you can see what they do before deciding whether to pay for them. Uploads also went up to 10 GB on every plan.

The Dubbing Editor Improved (Again)

• Fix the source transcript. Correct transcript mistakes directly and re-translate individual segments.
• Real speaker tools. Rename a speaker, remove one and merge their lines into another, add a speaker the diarizer missed, and play any speaker's reference clip inside the source video to work out who is who.
• Laughter and reactions survive. Vocal events are now layered under the speech they overlap instead of being thrown away.
• Export the dubbed audio on its own. Dub a video in multiple languages and export each one for uploading to YouTube.

Equations That Read Correctly

Math got another pass. LaTeX now reads out loud properly, so fractions, square roots, superscripts and subscripts, Greek letters, accents, and operator macros all come out the way you'd say them.

Also

• Voice Changer quality improved, with a large reduction in audible artifacts
• Settings reorganized into sidebar categories
• Audiobook exports larger than 4GB now work as expected
• Antivirus TLS inspection no longer blocks model downloads, VCP trusts your OS certificate store now
• Cloned voices without a transcript transcribe themselves instead of failing
• Runaway generations are handled in a better way