Download Whisper Model

Kon needs the {modelSize} model to transcribe speech.

{modelInfo[modelSize]?.size ?? "?"} · {modelInfo[modelSize]?.accuracy ?? "?"} accuracy · 100% offline

{#if downloading}

{progress}% · {formatBytes(downloaded)} / {formatBytes(total)}

{:else} {/if} {#if error}

{error}

{/if}

Models are cached locally. No data leaves your machine.