Preview a lightweight voice-style transfer, on-device.
Pick from Kokoro's built-in speakers.
Enter what the voice should say.
Download the rendered audio.
No — it uses Kokoro-82M's built-in speaker embeddings as style presets. Full one-shot cloning requires a much bigger model.
It gives you variety and consistency in a single voice per session, which is what most creators need.
No. All processing stays in your browser.
Paste text, pick a voice, listen or export a WAV. Offline.
Paste article text and listen hands-free while you work.
Extract text from a PDF and read it out with natural voices.
Turn an EPUB into chapter-by-chapter narrated WAV files.
Test SSML pitch, rate, pauses and emphasis in the browser.
Assign a voice per script line and export a merged WAV scene.
Hear a word in multiple accents side-by-side — US, UK and more.
Mix TTS narration with a background music bed via WebAudio.
Zro7