WikiTube

Legal

Privacy policy

Last updated September 14, 2026

What we collect

  • Account data. If you sign in with Google or GitHub, we receive your name, email address, and avatar from the provider. Authentication is handled by Supabase.
  • Usage data. We record which videos you analyze, like, or save to favorites, and count your daily analyses to enforce the free quota.
  • Crash reports. When the site itself fails in your browser, we send ourselves the error message, the URL of the WikiTube page you were on, and your browser's user agent. That is how we find out the site is broken — it is not possible to see a browser console from here, and a silent failure once went unnoticed for nineteen days. It contains nothing about other sites, and nothing you typed.
  • Analytics. We use Google Analytics for aggregate page-level usage. Article content, transcripts, and anything you write are never sent to it.

How we use it

Your data is used to operate the service: keeping you signed in, remembering your favorites and likes, and enforcing daily quotas. Video analyses are cached and shared — when you analyze a public YouTube video, the resulting article becomes readable by other visitors. The cache contains video content, not your identity.

What we don't do

  • We don't sell your personal data.
  • We don't track your browsing outside WikiTube.
  • The browser extension operates on YouTube and WikiTube pages, with a localhost address also supported for development. It does not request access to your general browsing history or read pages on unrelated sites.

Browser extension and sidebar

The extension reads YouTube video links, video identifiers, caption availability, and player state to display article shortcuts. It may check caption availability as video cards appear and prefetch captions when you hover over a shortcut. These requests omit your YouTube account cookies. When you open an article, its video identifier is sent to WikiTube; captions and video metadata are passed to WikiTube when needed to generate the article. Transcript content is processed by our AI service provider to produce the shared article.

Opening mode, sidebar layout, width, and language preferences are saved locally in your browser. Recent captions and availability results are cached temporarily in browser-session memory. The open sidebar's video and pending seek position are saved in the YouTube tab's session storage to restore it after a reload. While the sidebar is open, playback time is communicated to the embedded WikiTube reader to keep the article aligned with the video. Clicking a timestamp sends a seek request back to the YouTube player.

The sidebar loads the WikiTube website, so the account, usage, crash reporting, and analytics practices described above also apply there. Signing in uses a separate WikiTube window and browser cookies, including a separate session for the embedded reader. WikiTube sign-in tokens are not sent to YouTube or saved in extension storage. You can close the sidebar, change its preferences, sign out, or remove the extension at any time.

Third-party services

We rely on Supabase (database and authentication), OpenRouter (LLM analysis of video transcripts), and YouTube (video metadata and playback). Each processes data only as needed to provide the service.

Your choices

You can use WikiTube to read analyzed videos without an account. To delete your account and associated data, contact us at yunfei@iart.ai.