Language Island
For your Muse Library · Experimental

A little practice.
Your own island.

Turn things you want to say into a personal study library. Choose from 31 languages, hear them spoken, and come back when it’s time to practice.

Install with MuseView on GitHub ↗

Free app source. Runs in your own Muse environment. No Language Island account or paid speech API key.

01 / CREATE

Learn what you need.

Describe a situation. Choose a language. Get phrases with readings, explanations, and useful vocabulary, then save the ones you want.

02 / LISTEN

Speech on your machine.

Supertonic generates speech locally in 31 languages, including Spanish, French, Japanese, Korean, German, and Arabic. Recordings stay with your app for quick replay.

03 / REMEMBER

A rhythm for practice.

Review your cards with spaced repetition. Organize them into collections, favorite useful phrases, and export a backup.

Give your Muse an island.

One prompt. Your own Library app.

  1. 1. Copy the prompt

    The GitHub link gives Muse the app source, design, and setup instructions. No ZIP download needed.

  2. 2. Paste it into Muse

    Start a new conversation in your own Muse account. You’ll need access to fullstack Library apps and local execution.

  3. 3. Let Muse install it

    Allow its normal download requests. Muse builds your private app, installs the local voice model, and checks that it works.

Your installation prompt

Bring your library with you.

Already using Language Island?

  1. 1. Download your archive

    Export your library from Language Island Settings. Keep the original ZIP.

  2. 2. Attach it in Muse

    Install the Muse app, then attach your ZIP in your own Muse conversation.

  3. 3. Try five cards first

    Ask Muse to convert a sample, show what transferred, and check with you before importing the rest.

Get the migration prompt and instructions →

The formats differ. Audio and review progress may need adaptation; this assisted migration is not yet tested end to end. Your library does not transfer automatically.

An early release, ready to explore.

Muse does not currently offer a one-click installer for fullstack Library apps. Your Muse builds from the source reference, so setup may need adaptation.

A reset installation worked from one GitHub-link prompt in about 21 minutes, including checks, with one download approval. Spanish generation, saving, review, and cached playback were checked in the app. A second account and voice quality across languages remain untested. See all 31 supported languages. The interface and explanations are in English; voice quality varies by language.

Read the verification notes ↗

What to expect

  • Allow time for setup: our reset-install test took about 21 minutes, including checks.
  • About 396 MB of voice-model downloads; roughly 1.1 GB including the tested speech environment.
  • Muse usage limits still apply to generation and background tasks. This is not a fully offline app.
  • First playback can take time to prepare. Cached recordings replay without new synthesis.
  • Your cards and audio use your app’s private storage. No shared Language Island server.