VoxCPMTTS Examples

This directory contains the static project page and browser-side examples for Hangry Labs VoxCPMTTS.

Static Project Page

  • index.md integrates the gallery into the shared GitHub Pages website.
  • index.html remains the standalone gallery, with local styles in gallery.css.
  • voices.js embeds the generated sample manifest used by the page.
  • player.js renders the language picker and audio sample player.
  • background.js supports the page background.

Generated Audio

  • original_clone.mp3 is the project seed voice, generated from a randomized KokoroTTS voice and used consistently for clone examples.
  • assets/manifest.json indexes the generated samples.
  • assets/<language>/random/ contains 10 native-language voice-design samples.
  • assets/<language>/intro/ contains 3 translated project intro samples.
  • assets/<language>/clone/ contains 1 translated clone sample using original_clone.mp3.
  • assets/featured/ contains curated product demonstrations that pair generated audio with the UI used to create it.

SSML-H Documents

  • ssml-h/polish-launch-dialogue.ssml is an extended fixed-seed Polish launch sequence covering the countdown, emergency response, recovery, and launch.
  • ssml-h/english-launch-dialogue.ssml and ssml-h/german-launch-dialogue.ssml provide matching three-speaker launch scenarios for multilingual testing.

Regenerate and validate samples with the project maintenance scripts. After regeneration, rebuild voices.js from assets/manifest.json so the static page uses the latest sample metadata.

Preview the complete website with task pages and open http://127.0.0.1:9005/VoxCPMTTS/examples/. The overview is built from the root README, with a shared header and six interface languages.

For the standalone gallery, preview locally from the repository root:

python -m http.server 9000

Then open http://localhost:9000/examples/.