Privacy

Text you paste

Text entered into the tokenizer stays in your browser. Tokenization runs in a Web Worker on the page, and there is no code path that sends your input anywhere. Reloading the page clears it. Nothing you type is written to storage, logged, or associated with an identifier.

What the site stores

• No cookies set by this site.
• No analytics, advertising or third-party tracking scripts.
• No accounts, email addresses, contact details or lead capture.

What the hosting provider sees

The site is served by Cloudflare. Like any web host, Cloudflare processes network metadata such as IP address and user agent to deliver the page and to protect against abuse. That processing happens on Cloudflare's infrastructure, is governed by their own privacy documentation, and is outside this project's control.

Third-party assets

Tokenizer vocabularies are served as static files from this site's own domain. Fonts are bundled into the build and also served locally. The pages therefore load no scripts, fonts or data from third-party origins at runtime, so no third party receives your visit through page content.

Model and trademark naming

Model names such as BERT are used descriptively to identify which tokenizer is being inspected. They are trademarks of their respective owners, used here under nominative fair use. See the about page for details.

Changes

If this site ever gains a feature that stores or transmits anything, this page will be updated before that feature ships, and the change will be visible in the site's commit history.