New: refine your dub with a scenario script

You know exactly what your video is about. The AI transcribing it doesn't — at least, not yet.

Speech recognition is impressively good, but it still trips over the words that matter most to you: a product called TutDub, the library named vue-router, a colleague's surname, "3FA" instead of "three-F-A". When a term gets mangled in the transcript, the mangling carries straight into every translated version.

Today we're shipping Refinement using a scenario — a small toggle that lets you hand the AI the text script of your speech up front. It's available on Pro and Business plans.

How it works

On the upload screen, next to the new option, toggle it on and paste the script of what you say in the video.

That's the whole setup. There's no special format to learn, no XML, no "What should I write here?" screen — just plain text, up to 10,000 characters. A rough draft, a bullet list of key terms, the actual script, whatever you have.

Behind the scenes, when TutDub cleans and translates your transcript, it consults that text and uses it to do two things:

1. Refine the transcription — words that recognition misheard get corrected against what you actually said.
2. Lock down specific terms — product names, acronyms, technical jargon and proper nouns survive the translation instead of getting re-invented.

Where it helps the most

• Courses with a vocabulary. A "clean carve" is a stepper, not "clean curve". Tell the AI once, and every local version says it right.
• Screencasts full of UI text and filenames. The thing on screen has an exact name. Now the dub uses it.
• Named people and places. Proper nouns are the classic recognition trap. A one-line script fixes them across all languages at once.
• Formal or brand-toned material. If a term must keep its English name inside a Spanish sentence, say so and it will.

Do I need it?

No — and we want to be straight about that.

The base pipeline is not something you have to rescue. TutDub cleans fillers, clones your voice and translates on its own, well enough that the majority of uploads never touch this option. You can happily ignore it forever.

Refinement using a scenario is a precision feature, not a dependency. Use it when accuracy on specific words is the whole point of the video — branded trainings, certification prep, API docs, anything with terminology worth getting exactly right. When it's that kind of content, ten seconds of pasting can save a round of "this word sounds wrong" fixes on five languages.

The short version

• Light it up in the upload screen on Pro and Business.
• Paste the script of your speech — free-form, no formatting needed.
• TutDub refines the transcription and locks down your specific terms.
• Skip it anytime: the regular pipeline still works great on its own.

Try it on your next upload and hear the difference on your own words.