Privacy

Private beta · Updated September 26, 2026

Bagworth helps you identify handbags, estimate resale value, and keep a personal collection. This notice explains what the beta uses and where it goes.

Information you provide

You may provide bag photos, listing links, condition and year details, purchase prices, collection notes, an email address for sign-in, and optional profile information.

How the beta uses it

• Bag photos may be sent to Google Gemini or Anthropic to suggest a brand and model. Suggestions are not authentication.
• Pasted listing links are fetched by Bagworth to extract product details and may contribute non-private listing evidence.
• Purchase prices remain personal collection data and are not treated as market-comparable evidence.
• Anonymous collections are stored in the browser. Signed-in collections are stored with Supabase so they can load across devices.

Product analytics and support diagnostics

When enabled, Bagworth records page visits and steps such as requesting an estimate and saving a bag to understand usage and improve reliability. We use a random browser identifier, session identifier, broad traffic source and device type, event time, and your account identifier when signed in. Analytics does not include your email, photos, bag notes, purchase prices, or pasted listing URLs. Reports are restricted to approved admins.

To troubleshoot problems, we also record which step succeeded or failed (such as sign-in loading, photo preparation, estimation, saving, editing and collection reloads), timing, safe error categories, app release, and random operation and request identifiers. Approved admins can view account or session timelines. These diagnostic events contain no raw error messages, stack traces, file names, photos or bag content, and are deleted after 30 days on the next event ingestion. Delivery is best-effort and the browser choices below apply.

We also use Vercel Web Analytics for visitor and page-view totals, referral sources, and broad device and location statistics. We remove query strings and replace bag and closet names or identifiers in tracked paths with generic labels. Account, sign-in, and admin pages are excluded, and we do not send account identifiers or bag content to Vercel Web Analytics. The browser choices below apply to both analytics systems.

We respect browser Global Privacy Control and Do Not Track signals. You can also disable analytics for this browser below. This does not affect your collection or sign-in.

Bag improvement history

When enabled, we keep structured bag details, identification suggestions, your corrections, and estimate results for up to 90 days, even if you do not save a bag or sign in. This helps us find identification problems and understand which bags people check. Estimates and corrections are not verified sales and do not automatically change prices.

This history excludes account information, purchase prices, private notes and pasted URLs. It uses random bag-journey IDs and a browser deletion key. We respect Global Privacy Control and Do Not Track. Collection photos remain part of your private collection; using a photo for product improvement requires the separate, unchecked choice beside upload.

If you choose to contribute a photo, we keep a small image with camera metadata removed for up to 30 days. These images are private and are not displayed publicly. We do not automatically train models on them. Retention cleanup runs daily, so physical removal may take up to one additional day. History and images are excluded from reports after their retention window.

Deleting history also deletes contributed photos and prevents pending requests from restoring them. A hash of the deletion key remains only to enforce that choice. These controls leave your collection unchanged. Use them on each browser you used; clearing browser storage first loses its deletion key. If that happens, unlinked contributions expire through the retention policy. You can still request help through the person who invited you.

Pending estimate research

When a saved bag has no estimate, approved reviewers may research public listings using its confirmed brand, model, size, material, year and condition. Research requests and their results are private. Purchase prices, private notes, account details and label photos are excluded from search queries.

Using your main bag photo in Google Lens requires a separate, optional choice on that saved bag. If you allow it, Google receives a small copy with camera metadata removed to help identify the bag and research resale prices. You can withdraw permission on the bag page; changing the photo requires permission again. Withdrawal stops future exports, but cannot recall a copy already sent to Google. Public-listing research can continue without photo permission.

Research depends on available listings and reviewer availability. It does not guarantee an estimate or a completion time. Results remain dated and are only shown for the matching bag. Removing the saved bag also removes its research requests, photo permissions and results.

Sharing and retention

Bagworth uses service providers only to operate the beta, including Vercel, Supabase, and the configured image-identification provider. The beta does not sell personal information. Anonymous browser data remains until you remove it or clear the browser. Signed-in items remain until you delete them or request account removal.

Your choices

You can try Bagworth without signing in, remove bags from your collection, and avoid uploading photos or personal notes. During the private beta, send privacy or deletion requests through the person who invited you. A dedicated support address will be published before a broader public release.