Privacy notice

Scope and contact

This notice describes the public Mattevia image tool. The service currently offers free single-image processing within displayed limits; paid checkout is not publicly open. For privacy questions or requests, contact the Mattevia team at hello@mattevia.com.

Image processing and retention

Images are sent to our server-side processing service, not processed entirely in your browser. Cloudflare provides the website, database and R2 object storage; the image-processing service runs on Vultr infrastructure. Uploads are not used for model training.

Image access and retention follow the mode selected when the task is created. For existing single-image tasks in public-link mode, originals, results and comparison previews use hard-to-guess links: anyone who obtains a link can read the image without signing in. Originals are not deleted after processing, and scheduled cleanup for these files is not yet enabled. This does not guarantee permanent storage. Earlier private tasks retain their original owner-only access and 24-hour recovery window, including their original cleanup policy; expiry does not guarantee physical deletion at the exact expiry time.

Account batches have a 24-hour recovery and download window, shown with the batch. Basic batch and task metadata may remain in the database after that window. Scheduled cleanup removes the original images, results and previews associated with expired batches only after the relevant tasks have reached a terminal state. Cleanup is deferred for active tasks or tasks whose status is uncertain, so expiry is not a promise of immediate physical deletion. Deletion from storage does not immediately revoke copies already downloaded by users or held in CDN caches. This batch policy does not change the retention rules for existing single-image files described above.

Task history and recovery APIs still require task ownership. Public image links themselves do not require authentication; share them only with people you intend to give access. Storage credentials are never provided to the browser.

Replacing an image, resetting the workspace or closing the page does not itself delete a stored server result. Once the result reaches your browser, color changes, export and repeated downloads use the local image without uploading it again. Save the files you want to keep; account history is not permanent image storage.

Cookies and browser storage

Free processing uses signed anonymous cookies for usage controls. The necessary HttpOnly cookie cutout_visitor, limited to /api/cutout, identifies the anonymous owner for recovery. It lasts up to 24 hours; clearing it prevents recovery of that anonymous owner's tasks. Authentication uses session cookies. Language and appearance preferences may be stored locally. The cookie lifetime controls anonymous task recovery, not file retention, and does not revoke an already obtained public image link.

In the single-image workspace, for recovery, sessionStorage keeps the current request identifier and up to 10 recent task identifiers, their times and a storage-type indicator. It does not hold image binaries or signed image URLs. Resetting the workspace preserves these entries; clearing site storage removes browser entries, not server records.

The batch workspace uses IndexedDB in the current browser to cache draft files that have not yet been uploaded. Uploaded account batches are recovered from the server rather than relying on this draft cache. An entry point remains available for older batches stored locally under the previous browser-storage approach. Those older local records retain their 24-hour recovery window, with expired-record cleanup performed when that storage is accessed rather than at an exact deletion time. Clearing site data removes local drafts and older browser copies, not server-side batches or images.

Accounts and operational data

Free processing does not require an account. If you create an account, the service stores account and session information and relevant task history in its database. Image binaries are stored separately from account records. Closing the browser does not delete account data or history, and image expiry does not automatically delete all database records or backups.

Operational logs may contain request and error information used to maintain the service and investigate failures. When product analytics is enabled, processing and download events use predefined action names. These product events do not include images, filenames, task identifiers or storage keys. Analytics availability depends on deployment configuration.

Requests about your data

Email hello@mattevia.com to ask about stored data or request account deletion. We may need to verify account or task ownership. There is currently no self-service account deletion or data-export flow. Please do not send passwords or sign-in codes.

Contacting us does not extend image recovery or guarantee restoration of expired files. If you choose to send an email or an attachment, it is handled through our support mailbox separately from the image tool; the tool's image retention rules do not apply to that correspondence.

Trial and payment records

When trial or paid processing is enabled, we keep pseudonymous trial identifiers, account identifiers, task usage, orders and payment-review records to enforce allowances, prevent duplicate charges and reconcile refunds. Anonymous trial limits can be shared by a network; we do not store the raw network address in the trial ledger. Payment credentials are handled by the payment provider. Refund reviews can pause new processing while retaining already completed results under the rules above. Contact us about access or deletion requests; records needed for outstanding payments, fraud prevention or legal obligations may need to be retained.