All insights

Claude + ElevenLabs → narrated MP4, in four languages

The same Acme contact flow as the Polly demo, now in four languages from one source script. Claude translates each step — narration, captions and UI labels. ElevenLabs (George, a British voice on the multilingual model) speaks every language. ffmpeg stitches one MP4 per locale, each with its flag dropped into the title bar in post.

How it is made

  1. Capture. Playwright drives the app, applies spotlight CSS and screenshots clean 1920×1080 frames. Shared across every language.
  2. Translate. Claude translates each step — narration, captions, UI labels — into French, Spanish and Japanese.
  3. Narrate. ElevenLabs synthesises one MP3 per step in the George voice. A single multilingual model covers all four languages.
  4. Stitch. ffmpeg composites each still, the title and caption bars and a per-locale flag, then concatenates one MP4 per language.

One script, one set of screenshots, four narrated walkthroughs. Localising the product demo stops being a project and becomes a build step.

Talk to us about this

Contact

Talk to us before your next AI investment.

We run this work in our own business every day. If you are planning to run it in yours, the first conversation is diagnostic, not a pitch.