No accounts. No cookies. No ads. We collect anonymous, aggregate usage statistics and nothing else. Your favorites and preferences live in your browser and never leave your device.
We use Vercel Web Analytics, a cookieless analytics service. It records page views and a small set of anonymous feature-usage events (for example “a download button was clicked” or “a feed failed to load”). Visitors are identified only by a hash that is discarded daily — there is no persistent identifier, no cross-site tracking, and your IP address is not stored.
Usage events never include what you searched for or personal identifiers of any kind.
Favorites, content preferences, consent choices, and small one-time UI flags (such as “the install hint was already shown”) are stored in your browser’s localStorage. We never see them. Clearing your browser data removes them.
The multi-reddit builder’s smart search uses an AI model (xAI’s Grok, accessed through Vercel AI Gateway) to turn your search text into subreddit suggestions. Only the text you type into the builder — plus the names of subreddits it previously suggested in the same session — is sent, from our server. No cookie, IP address, or any identifier linking the request to you is attached.
Vercel AI Gateway does not retain prompts. xAI stores API requests for up to 30 days for abuse monitoring, then deletes them automatically, and does not use them to train models. If you prefer not to share search text with an AI provider, the preset combinations on the same screen work without it.
All media is loaded from Reddit’s servers. Your browser talks to Reddit’s content delivery network directly when viewing media, subject to Reddit’s privacy policy.
Like virtually every website, our hosting provider (Vercel) keeps short-lived operational request logs used for debugging and abuse prevention. They are automatically deleted after a retention period. We also record anonymous media-failure categories (for example “a video from an external host failed to load”) — coarse labels only, never the specific content or anything identifying you — so we can fix broken media.
Peekstr is an independent project. If you have questions about this policy, reach out via the support link on the homepage.