Site information
Privacy and data use
What happens when you read WebML or try an experiment.
WebML is a personal browser AI lab maintained by Omri Beladev. You can read the articles without an account. For questions about this page, email hello@webml.ai.
Analytics and advertising
The site loads Google Analytics, PostHog, and Vercel Analytics to understand visits and site usage. Those providers may receive information such as your page URL, browser, device, and network address. Their own services govern how they process that information. The site also loads the Google AdSense script, which may use cookies or similar technologies for advertising if ads are enabled. See how Google uses information from sites that use its services and PostHog's privacy policy.
Local AI experiments
For the Playground and voice agent, model inference is designed to run in your browser. Starting an experiment can download model files from third-party hosts and store them in browser caches. The voice agent asks before loading its large models and requests microphone access only when you start listening. The site does not offer an offline page: a cached model does not guarantee the website itself will open without a connection.
Analytics still run on experiment pages. Avoid entering sensitive information into a demo, and review the linked source code if you need to assess a particular data flow before use.
Contact and control
If you email us, your message and address are used to respond. You can clear this site's stored model files and site data in your browser settings. Browser privacy controls and extensions may limit analytics or advertising requests.
For the technical work behind the site, see the articles and lab.