How to use it
Enter up to 1,500 characters, choose a preset and pace, confirm responsible use, then select Generate speech. Listen or download the result.
Text to speech studio
Give your words
a voice.
Turn a thought, a story, or a lesson into clear synthetic speech. Choose a preset, set the pace, and listen.
Start creatingThe studio
Make something sound good.
Generate speech on this device. Your text is not sent to a Clearvoice server.
0 / 1,500 characters
0.8 slower · 1.0 natural · 1.2 faster
0.81.2
Generated audio
Your audio will appear here after generation. The first request downloads the speech engine and model to this browser.
Download audio
AI-generated speech can be mistaken for real audio. Do not use it to impersonate people or mislead listeners.
About Clearvoice
Speech, made simple.
Clearvoice turns written words into listenable audio with the open-weight Kokoro model and preset voices. Voice labels describe documented accent and vocal traits; they do not verify age or identity.
Privacy in brief
Speech generation runs on this device. Model and app files are downloaded from their providers; network providers receive ordinary connection metadata.
Responsible use
Do not impersonate someone or use generated audio to mislead, harass, or defraud. Follow applicable disclosure rules.
Help, privacy & use
How this app handles your text and audio
The app generates speech in the browser and does not send your text or audio to a Clearvoice server. The speech model is downloaded from Hugging Face and cached in your browser. Your browser makes model-file requests to Hugging Face; your network provider and hosting provider can see ordinary connection metadata. Browser storage may be cleared or evicted at any time.
Acceptable use
Do not use this app for impersonation, fraud, harassment, deceptive endorsements, or other harmful synthetic audio. Respect people's rights and the disclosure rules that apply to your platform and location. This reminder is not technical enforcement.
About the speech model and limitations
Clearvoice is an independent interface and is not affiliated with the Kokoro model authors. Speech is AI-generated and may mispronounce words. Preset voices are not voice cloning or verified representations of real people. The first model download is large and requires internet access; the browser may cache it afterwards. Performance depends on the visitor's browser, device, memory, and processor.