PassionFeed

Your visual world. Your choice.

PassionFeed is a visual discovery app with optional Google sign-in. Every visitor has a private server-saved library. Google sign-in makes it available across browsers.

Your private server library

Your interests, preferences, feed timelines, photo metadata, review batches, Collection, Loves, hides and seen-image history are saved in PassionFeed’s Cloudflare database. Discovered photos are stored as source URLs and metadata. Pixabay photos are copied to our server only after you add them to an interest feed. Candidate previews are not copied. Stored Pixabay files are deleted when no accepted interest feed uses them; hiding a photo removes your memberships, while another user’s accepted copy may remain. Dismissing an unaccepted candidate stores no image file. Without Google sign-in, a random HttpOnly cookie identifies your temporary user; the database stores only its credential hash. The cookie lasts up to one year. Clearing cookies loses access to that temporary library and does not delete its server records.

Photos you upload

Signed-in users can create My photos interests containing only their own uploads. Photos are prepared in your browser as high-quality JPEGs, with location metadata removed, and stored privately in Cloudflare R2. Only your signed-in account can view them. Uploads are not used for discovery or sent to AI providers. Hiding a photo or removing its interest revokes access immediately; unreferenced files are deleted, with failed deletions retried on later visits. Your account has a 500 MB upload allowance. These interests never collect external images or generate AI images.

What stays in your browser

The open page keeps a working view in memory. New libraries are not saved in IndexedDB or local storage. Existing browser libraries are imported once, and their local copies are removed after the server confirms the import. Old copies belonging to a different account remain until that account signs in to import them. A connection is required to load and save your library. If saving fails, keep the page open and retry; unconfirmed changes can be lost if you leave.

Optional Google sign-in

Signing in shares your Google account identifier, name and verified email with PassionFeed. We request only basic sign-in scopes, with no access to Google Drive, Photos or other Google content. Your temporary library is merged into your Google account on the server, combining matching interests. Its temporary credential is then revoked. Account data remains until you change it; clearing browser data does not delete the server copy. Signing out starts a separate temporary library. A private sign-in session cookie lasts up to 30 days and is revoked when you sign out. Google access and refresh tokens are not retained. Daily generation limits apply to server users and temporary usage transfers when you sign in.

What is processed on the server

When you view an interest, its phrase, preferences, source URLs, image IDs and limited Loved-photo title summaries are sent to PassionFeed to find and assess future discoveries. Collection-run results are cached for recovery and removed after seven days through housekeeping when the app is accessed. Shared image URLs, source metadata and assessments may be retained to avoid repeated discovery and judging. AI-generated images are stored without automatic expiry so Loved images remain available. Other discovered image files remain on their original hosts. Accepted Pixabay photos are the exception described above. Failed file deletions are queued and retried when the app is accessed.

Usage analytics

We send page visits and successful feed and film creation actions to our private Lift Dashboard. Records include a random visitor identifier, account identifier and signed-in name when available, event time, page path, referring website, IP address and approximate city, region and country. Prompts, photo bytes and film contents are not included. A separate HttpOnly visitor cookie lasts up to one year. Analytics help us understand app usage and exclude owner testing.

External services

Discovery may use Unsplash, Pixabay, Pexels and Lummi when their APIs are connected, Flickr directly or through Openverse, museum collections, public web/image search and user-supplied sources. The browser loads discovered images directly from their hosts, which receive ordinary connection information such as your IP address. Unsplash hotlinks and tracked download requests also provide photo interaction information under Unsplash’s privacy policy.

OpenRouter and its model providers process interest/preferences text, limited feedback summaries and remote image URLs for visual review and optional image generation. See OpenRouter’s privacy policy. Cloudflare hosts PassionFeed and processes connection/security information. Hashed IP addresses are used for abuse limits and a testing exemption. Provider-cost and quota records are retained for operations.

Recommendations and control

Loved-photo title summaries are designed to influence future judging when it is enabled; LLM judging is currently disabled. Time spent on an image, scroll speed and other passive engagement signals are not used for recommendation scoring. Seen-image IDs help reduce repeated discoveries. Hiding a photo removes it from all your interests, review batches, Collection and cached collection results. Its ID remains on your account’s exclusion list to prevent it from being offered again. Hide titles are not kept as recommendation feedback. You can remove interests, unlove Collection items and clear the browser’s site data. Removing an interest removes all associated photos from your account, including Collection and other interests. Stored files are deleted when no other account uses them. Shared card images are prepared temporarily in your browser and are not uploaded to PassionFeed.

Return to PassionFeed