Your diary is personal.

Log My Dish stores your account email, a password verifier if you use email sign-in, preferences, diary entries, uploaded label images, favorites and drafts in its private database and image storage. Saved meals remain until you delete them.

If you choose Google or Apple, that provider authenticates you and shares a stable account identifier and verified email with Log My Dish. Apple’s private relay email is supported. We do not receive your Google or Apple password, request your contacts, or send your diary to the sign-in provider. Linked identities can be managed in Settings. Deleting your Log My Dish account removes its linked identities; provider-side permissions can also be removed in your Google or Apple account settings.

When you request AI analysis, your meal description or label image is sent to the server’s configured AI provider. Conversational revisions include the relevant prior meal. The OpenAI adapter requests that responses are not stored using store:false; provider abuse-monitoring and retention policies still apply. Optional USDA lookup sends food names to FoodData Central.

FatSecret searches send your search terms and selected food IDs to FatSecret. When you enable AI search improvements or request customization analysis, your context and relevant food/portion descriptions go to the configured AI provider. Personal recipes retain your ingredient inputs, selected quantities, yield and food IDs. FatSecret names, brands, serving definitions and database nutrients are not stored as a server catalog or permanent diary snapshot. These values are retrieved again when needed. Provider content is kept in the active view only; recovered device drafts retain references and your own edits and require a fresh lookup.

OAuth connects ChatGPT to the same account with read and/or write scopes. You can revoke connections, remove uploaded images, export your diary or delete all account data in Settings. Device-local unsaved input uses IndexedDB so a failed request does not erase it. The app does not cache authenticated API responses or raw diary data in its service-worker cache.

The application does not log raw descriptions, photos or credentials by default. Hosting infrastructure may retain routine request metadata. Nutrition estimates can be inaccurate; the review step lets you correct them.