Privacy Policy
Last updated: October 29, 2026
Overview
Kokoro Web (kokoroweb.app) provides in-browser text-to-speech. We value your privacy and this policy explains how we collect, use, and protect your information.
Information Collection
Text Input and Generated Speech
- The text you enter is passed to a worker in your browser to generate speech. The application does not upload that text or the generated audio to our servers for synthesis.
- Speech synthesis runs on your device using WebAssembly or WebGPU where available. This tool does not record microphone audio or transcribe recordings.
- Draft text and generated audio are held in the page session. Save your text and download any WAV files you want to keep before leaving or reloading the page. Downloaded files are stored on your device under your control.
Browser Storage and Downloads
- The browser may cache model and voice files to avoid repeated downloads. The application requests persistent storage where supported, but browsers can still limit or clear cached data.
- Clearing this site’s browser data removes its local caches and may require another model download. It does not delete audio files you have already downloaded to your device.
Information Use
- Provide and maintain the browser-based text-to-speech and related site features.
- Cache model files in your browser where supported.
- Measure site usage and advertising performance through the services described below.
- We do not sell your text, audio, or personal information.
Third-Party Services
- Our hosting, content delivery networks, and model-file providers receive the network information needed to deliver requested files, such as your IP address, browser details, and requested URL.
- This site uses Google Analytics. Google may process page-view, device, browser, and event information to report site usage. See how Google uses information from sites that use its services.
- Model downloads may contact Hugging Face or a configured backup file host. Pre-recorded voice samples are served as audio files. These requests disclose network information to the serving provider; local speech synthesis does not mean that visiting the website makes no network requests.
- Japanese pronunciation processing downloads a version-pinned Open JTalk worker, WebAssembly module, dictionary, and related files from jsDelivr on first use. These requests disclose network delivery information to that provider; the entered text is processed in local browser workers and is not sent to a pronunciation server. Mandarin processing loads a local bundled pronunciation dictionary when needed.
Google AdSense and Advertising
This site uses Google AdSense for advertising. When advertising code runs, Google and participating advertising providers may place or read cookies and similar storage, use web beacons, and receive information such as your IP address, browser or device details, page URL, and advertising identifiers. They may use this information to serve and measure ads, limit repeated impressions, and detect invalid activity. The technologies and data used depend on the ads and consent settings in effect.
Third-party vendors, including Google, use cookies to serve ads based on your prior visits to this site or other websites. Google’s advertising cookies allow Google and its partners to serve ads based on those visits.
You can opt out of personalized advertising or change your ad personalization settings in Google Ads Settings. For information about Google's advertising cookies, see How AdSense uses cookies. Other participating providers may offer their own privacy and opt-out controls.
Cookies
Kokoro Web may cache model files in browser storage. Google Analytics and AdSense may use cookies or similar technologies as described above. You can use browser controls to block or delete cookies and clear site storage; doing so may affect site functionality and require model files to be downloaded again. Google ad personalization controls are linked above.
Data Security
- Speech synthesis happens in your browser; the application does not send input text or generated audio as analytics or advertising events.
- If you contact us by email, we receive the address, message, and any files you choose to send so we can respond. Avoid including sensitive text or audio that is not needed for your request.
- Website is served over HTTPS.
Children's Privacy
Our service is not specifically directed to children under 13. We do not knowingly collect personal information from children under 13.
Policy Changes
We may update this privacy policy from time to time. Any significant changes will be communicated by posting the new privacy policy on our website.
Contact Us
Questions? Visit our contact page or email [email protected].