ORVIO
ON-DEVICE VOICE TOOL

Register a voice.
Read text aloud.

Record or import audio. Enter a sentence.
Create short audio on your phone.

iOS release in preparation

Planned as a free initial release. No ads or in-app purchases.

The orange ORVIO app icon
Register a voice

Edit a script and record, or import WAV, M4A or MP3 audio. 3–15 seconds, up to 20 MB. Prepare while the model downloads.

Generate on-device

Works offline after model setup. Recordings and text are not sent to a speech-generation server.

Export WAV audio

Listen, save and share. Generated audio stays in your local history.

Before you start

iOS
iOS 17 or later. Up to 80 characters and approximately 11 seconds per generation. Keep the app open while processing.
Initial download
Approximately 884 MB from Hugging Face. Wi-Fi and at least approximately 1 GB of available storage recommended.
Speech languages
Japanese, English, Chinese, Korean, German, French, Russian, Portuguese, Spanish and Italian. Explicit language selection or automatic detection. Text is not translated.
Interface
Japanese and English, following your system or your selection in Settings.

Generation can take minutes depending on the phone and text. Quality, accent and voice similarity vary by language and recording; equal quality and speed across all devices or languages are not guaranteed.

Use only your own voice and disclose that the audio is synthetic when sharing or publishing it.

Uses a quantized Qwen3-TTS model and qwen3-tts.cpp. ORVIO is an independent app, not an official Qwen, Alibaba Cloud or Hugging Face app. Android is also in development.