Privacy Policy

Last updated: August 1, 2026

The Short Version

OfflineTTS is designed so that speech synthesis and transcription inference run in your browser. The application does not upload your source audio or generated audio to OfflineTTS for processing. One content exception is documented below: non-English Kokoro sends the text you enter to api.offlinetts.com for phonemization before your browser generates audio. The website also uses hosting, analytics, model-delivery, and optional affiliate services that receive ordinary web requests under their own policies.

Data We Collect

OfflineTTS does not currently require an account, payment profile, or API key. The browser application is not designed to send English TTS text, STT recordings, generated audio, or exported files to OfflineTTS for speech inference. Files you select are processed in the browser unless a tool explicitly describes another data path.

We use Google Analytics (GA4) and Microsoft Clarity to understand website activity and diagnose page-level usability problems. Depending on configuration, browser settings, consent controls, and the provider, these services may process page URLs, approximate location derived from an IP address, device and browser details, referral information, interaction events, session information, and cookie or similar identifiers. This is usage data; it is distinct from the text and audio content processed by the speech tools. See Third-Party Services for provider and opt-out information.

Phonemization (Non-English TTS)

Kokoro's browser library natively supports English phonemization — English voices work fully offline. For Japanese, Chinese, Spanish, French, Hindi, Italian, and Portuguese, phonemization requires specialized models (misaki, espeak-ng) that are too large to run in a browser.

  • What is sent: Your plain text is sent to api.offlinetts.com and converted to a phoneme string
  • What is received: A phoneme string returned to your browser, where audio synthesis happens locally
  • What the application does not submit to this endpoint: Source audio, generated audio, account details, or payment information
  • Retention: The phonemization application is configured not to persist submitted text as application content; infrastructure may still process standard security and request metadata
  • English Kokoro: Does not use this phonemization endpoint after required assets are available

How It Works

  • TTS models and STT models (approximately 120–590MB for Whisper Tiny through Small) are downloaded to your browser on first use and cached in IndexedDB
  • All speech generation and transcription happens on your device using WebGPU or WebAssembly
  • English Kokoro does not send text to the phonemization endpoint after the model and application assets are available
  • Non-English TTS voices send plain text to our phonemization server, which returns IPA phoneme strings (pronunciation data). Audio synthesis still happens locally
  • TTS audio is generated in-memory and can be downloaded as WAV files
  • STT audio is decoded in a background Web Worker and transcribed locally
  • The OfflineTTS application does not upload source or generated audio for speech inference

Model Files

AI model files (TTS and STT) are served from Cloudflare R2 (CDN) and Hugging Face. When you load a tool, your browser downloads the required model files. A model host receives a normal HTTPS request and may process IP address, request headers, timestamps, and security logs under its policy. Model downloads are not the site's only network activity: the page itself, analytics, fonts or other static assets, non-English Kokoro phonemization, and links you choose to follow can also create requests.

Cookies and Browser Storage

GA4 and Microsoft Clarity may use cookies or comparable browser storage for measurement, subject to their configuration and your browser or consent choices. OfflineTTS also uses browser storage such as IndexedDB, OPFS, Cache Storage, or local storage for model caching, worker communication, preferences, and offline performance. Local tool storage stays in that browser profile until it expires, the application replaces it, or you clear it.

Delete Local Data

To remove cached models, tool preferences, and other local application data, use your browser's site-data controls for offlinetts.com. Clearing only normal history may not remove IndexedDB, Cache Storage, or OPFS, so choose the control that deletes site storage. The next visit may need to download models again. Audio, transcripts, subtitles, or documents you exported are ordinary files on your device and must be deleted through your operating system or file manager.

You can limit analytics with browser privacy settings, content blockers, provider opt-outs, or applicable consent choices. The Google Analytics opt-out link is provided below. Those controls do not delete exported files and may not erase data already handled by a third party; consult that provider's policy for its request process.

Third-Party Services

We use the following third-party services:

  • Google Analytics (GA4): We use GA4 for aggregate site measurement such as pages visited, referrals, device type, and interaction events. Google handles that data under its policies. You can use the Google Analytics Opt-out Browser Add-on.
  • Microsoft Clarity: We use Clarity for interaction analytics such as heatmaps and session-replay-style diagnostics. Microsoft describes what Clarity collects, masks, retains, and links to cookies in its privacy documentation.
  • Cloudflare: The site and some model assets use Cloudflare infrastructure, which processes normal network and security information such as IP addresses, requested URLs, headers, and timestamps.
  • Hugging Face and other model sources: A browser may request model or runtime assets from an upstream host. That host handles the network request under its own policy.

OfflineTTS does not intentionally place TTS text or audio content into analytics events. Nevertheless, avoid putting confidential content into a URL, page title, support message, or another field that a website service could ordinarily observe. The non-English Kokoro content path is the separate phonemization exception above.

Affiliate Links

Some clearly marked links are affiliate links. Clicking one leaves OfflineTTS and allows the destination or affiliate network to receive the normal web request and use attribution cookies or similar technology under its own policy. Any account, content, purchase, or payment information you provide after leaving the site is handled by that third party. See the Affiliate & Trademark Disclosure for where paid links appear and how recommendations are reviewed.

Changes

If data handling changes materially, we will update this page and its “Last updated” date. Our product goal is to minimize server handling of speech content, keep audio inference in the browser, and document exceptions plainly enough that you can make an informed choice.

Contact

Questions about privacy? Contact us at contact@offlinetts.com.