VLPremiere Pro 0.11.86: A Second Voice for Narration

VLPremiere Pro 0.11.86 is out, released November 2, 2026, and it is a Voiceover Studio release. Its headline is choice: a second voice engine to narrate with, closer control over pauses, phrasing and pace, and finished voiceovers that land on your timeline as they are made. It also includes everything in 0.11.83. The full notes are on the 0.11.86 release page.

A second voice engine to choose from

The new option is Qwen3-TTS. You install Qwen3-TTS 1.7B (3.1 GB) from Settings > AI components and choose it under Voice engine on the Voice step, or set it for every new voiceover in Settings. Chatterbox stays built in, so this is an addition, not a replacement.

For professional editors, a second engine means a second read of the same script to compare before committing a film to one voice. For beginners, it is simpler: if the built-in voice does not suit a project, there is now another voice to try inside the same studio, without leaving the editor to find a separate narration tool. The release notes ask for one sensible check: listen to your first voiceover on Qwen3-TTS before you publish it, and tell the Community how it compares with Chatterbox for your voice.

Pauses, phrasing and pace you control

Three new cards on the Finish step shape how a part is read. Where the voice stops speaks a part in the pieces a reader would pause between. Pauses at your own marks adds real silences after your commas, dashes and ellipses. Pace on long parts sets how slowly long parts are read.

Pros gain phrasing that follows the script they wrote, rather than a flat read they have to accept or re-record. Beginners gain the same benefit without learning audio editing first: write the pauses into the text, and the voice respects them. Narration-led work — documentaries, explainers and courses — lives or dies on that pacing.

Finished voiceovers land where you work

On the Finish step, Put it on the open timeline when it is made now places the finished voiceover at the playhead with a marker for every part. It is on by default and remembered. The finished card also gains Show in Finder, and the file is always copied into the Voiceover folder beside your project and its Voiceover bin. On Preview & adjust, Save script .txt saves the voiceover's script as a text file.

That removes repeated friction: no hunting for a finished file, and a plain text script for records or captions work. The guide walks through the studio steps.

Models and the voice engine stay current, safely

AI models now stay on the versions tested with the app. Every model installs the exact version tested with your app; a newer one is offered as Update in Settings > AI components only when a new app brings it tested, and Go back to the previous version switches back at once.

The voice engine itself now updates too. It was installed once and never replaced, so the voice improvements in 0.11.83 may not have reached your studio. From 0.11.86 the app updates it when it opens — never while the studio is working. Voiceover problems are also logged now, so support can see why something stopped. Help: support.

Fixes that remove dead ends

Three fixes matter in daily use. When the whole-file voice check failed, Make the final voiceover saved nothing and offered no way out; the card now says why and offers Listen, Save Anyway and Show in Finder. When the studio folder was reached through a link, the finished voiceover could be left in the studio's hidden job folder instead of beside your project; it now always lands beside the project. And for Russian stresses, picking a stress from the menu for a word that is two different words (за́мок / замо́к) now changes that one place only, and a stress typed for a rare name is no longer refused.

Each fix is about trust: a failed check explains itself, a finished file is where the project expects it, and a stress correction applies exactly where you put it.

What this means for pros and for beginners

For pros, 0.11.86 is a narration quality and reliability update: two engines to cast from, phrasing controls on the Finish step, timeline placement by default, and tested model versions with a one-click way back. For beginners, the same update reads as ease: choose a voice, mark your pauses in the script, and find the finished narration already on the timeline, with markers for every part. No extra app, and no per-minute cloud step is described in the notes — the engine installs on the Mac from AI components, in line with how VLPremiere Pro runs its built-in AI on your machine.

Getting the update

Update from the app (VLPremiere Pro > Check for Updates…) or install it over your current copy: projects, settings and your licence stay in place, per the release notes. VLPremiere Pro is available by subscription at vlpremier.com, with a free trial of the full Pro app for 3 days and no credit card required. There are two plans — Creator for individual creators and Pro for the full advanced toolset including Voiceover Studio and on-device AI — you can cancel anytime, and your projects stay yours even if a subscription ends.