Narrate, translate and click through a product demo without recording it again

You recorded a good demo. It shows the feature, the cuts are tight and the secrets are blurred. Then the requests start. Sales wants it narrated, because nobody reads captions on a call. The team in Munich wants it in German. Marketing wants a version on the website that visitors can click through at their own pace.

Each of those used to mean another recording, or another tool. With DemoTake 1.3 they're edits to the take you already have. This post walks through all three on one take, and what stays on your Mac along the way.

#The take we start from

A five-scene take of a project tool: sign in, create a project, invite a teammate, set a due date, see the board. Every scene has a short description from when the take was planned, like "Create a project in one click". A few scenes have captions. That's all the voiceover, the translations and the click-through demo need.

Aurora walkthrough5 scenes · 0:52

1. 1Sign in✓ Dashboard visible0:07Recorded
2. 2Create a project✓ Project page open0:11Recorded
3. 3Invite a teammate✓ Invite sent0:12Recorded
4. 4Set a due date✓ Date on the card0:09Recorded
5. 5See the board✓ Board with 3 columns0:13Recorded

#Give it a voice

In the editor's Captions tab, the new Voiceover group has one button: Write Voiceover Script. DemoTake writes one short line per scene from its description. With DemoTake AI set up, the lines are written to fit: each one short enough to be said in its scene at a natural pace. You read them, and change any word you like. They're plain text fields.

Then you pick a voice. There are two kinds:

• On this Mac: the voices that come with macOS. They're free, they work offline, and nothing leaves your Mac. Add an Enhanced or Premium voice in System Settings for the most natural sound.
• DemoTake AI voices: natural voices that speak every language, on Director, Team and Enterprise, or with your own AI provider key. They use voice minutes from your plan, and DemoTake shows how many are left before you spend any.

Click Generate Voice and the lines land on the take, timed to their scenes. They start where their scene starts, so when you trim a scene later, its line moves with it.

#When a line doesn't fit

The one rule DemoTake never breaks: it doesn't speed up a voice to make it fit. A line that runs past its scene says so, in seconds: "Runs 1.4 s past the end of the scene." You can shorten the line, or click Hold Scenes to Fit, which holds each scene's last frame until its line ends. Both are ordinary edits, and both undo.

#The rest of the mix

If the take has music, it dips under the voice and comes back after it. Mute microphone under voiceover silences whatever the microphone picked up while you recorded, so you don't hear yourself under the new voice. And Captions follow voiceover replaces the captions with the voiceover itself, one caption per sentence, timed with the voice.

#Subtitles come free

Once a take has a voiceover, the Export sheet can add SRT and WebVTT files, a transcript, and a subtitle track inside the MP4. All of it is timed to the edited video, after every cut and hold. Share the video as a link and the subtitles show up in the player and in embeds.

#Make it in German and Japanese

Under Voiceover sits Languages. Click Add Language, pick German, then Japanese. DemoTake notes the take's own language (English here) from the take itself, never from your Mac's settings.

BEFORE

Re-record per language

• A new recording for every language
• Edits and blurs done again
• Voices recorded in a studio

WITH 1.3

One take in three languages

• Same footage, edits and blurs
• Captions, voiceover and titles translated
• A video per language from one Export

Translate with offers two ways:

• DemoTake AI translates every language. Only the words are sent. Your app's name, numbers, keyboard shortcuts, links and anything Secret Guard flags are swapped out first and put back on your Mac.
• On this Mac uses the translation built into macOS 26, for the languages you've installed. Nothing leaves the Mac at all.

Captions, voiceover lines, scene titles and descriptions, title cards and the tips of your click-through demo are translated. Then Review Translations… shows every line next to its translation. Labels say which ones changed in the original since they were translated, and which ones you edited yourself. Your edits stay when you translate again.

#Names that must not change

Your product is called Aurora in every language, and your app's buttons might stay in English too. Put those words in the Glossary and they come back exactly as written. Or give a term one fixed translation per language, so "Projects" is always "Projekte" in German.

#The text inside the app

The app you recorded has its own text, and you can't change those pixels. Text on Screen reads a few frames of every scene, on your Mac, and picks the words that matter: large text and text near a click. Each picked word gets a small translated label beside it in that language's video. The label follows the camera and never covers the original. Text under a blur is never read.

#A voice for every language

Each language gets its own voiceover, with a voice for that language. Translated lines can run longer than the English ones. When they do, Hold Scenes to Fit holds scenes in that language's video only; the English one keeps its timing.

Preview in shows the whole take in German or Japanese right in the editor, before you export anything. In the Export sheet, check the languages you want and every format comes out once per language, with the language at the end of the file name (-de.mp4) and German subtitles and a German transcript next to it.

#Let visitors click through it

Some people won't press play. For them, open the new Demo tab and click Make Interactive Demo.

1. Scene titlefrom its description
2. →
3. Clickhotspot where you clicked
4. →
5. Typingthe typed text on screen
6. →
7. Resultthe checked screen

Each click becomes a step: a picture of the screen, a pulsing hotspot on the button you clicked, and a tip. The tips come from what the take already says, the caption on screen or the voiceover, so the work you did above pays off twice. Drag a hotspot to adjust it, rewrite a tip, merge two steps, or let a step move on by itself.

The pictures are drawn from the edited video only. Every blur is in them, and typing you masked never shows: a password step simply says "Enter your password."

Preview plays the demo exactly as visitors will see it. Viewers click the hotspots, use the arrow keys or swipe on a phone, turn the voiceover on, and switch between English, German and Japanese.

Then either:

• Export it as a website: a folder whose index.html works offline, plus a ZIP. Put it on any host.
• Share… it as a link, and copy the embed code into your site, your docs or Notion. On a subscription, add a button at the end ("Book a demo") and see how far viewers get, step by step. Analytics use no cookies and never store IP addresses.

#What stays on your Mac

DemoTake AI voices and DemoTake AI translation are there when you want them, and they only receive the text they work on. Never your recording.

#Ask your agent

Everything above is also a tool call. If an agent recorded the take, it can finish the job:

Narrate the Aurora walkthrough with a Mac voice, translate it into German and Japanese, and make a click-through demo I can embed on the pricing page.

That's voiceover_script, voiceover_generate, take_languages_set, translate_take, interactive_build and interactive_share, in that order.

#Try it

Download DemoTake to try it on a take of your own. The details are in the docs: Voiceover and subtitles, Your demo in other languages and Interactive demos. For sharing, embeds and analytics, see Embeds, calls to action and analytics.