Why should browsers upload files directly to object storage?

Why should browsers upload files directly to object storage?

Direct browser uploads send file bytes from the user's device to object storage instead of relaying them through the application server. This can reduce server bandwidth and processing pressure, especially when many guests upload large photographs at the same time.

Separate the data transfer from the product rules

The application server still decides who may upload and where the file belongs. It authenticates the guest, checks wedding membership and quotas, creates an upload record, and returns temporary storage permission.

The browser then transfers the file to the approved destination. Cloudflare's upload guide describes this direct-upload pattern. The browser should never receive the long-lived secret key used by the backend.

Design around an incomplete upload

Show progress and allow the user to recover from connection failures. A failed transfer should not appear as a finished gallery item. If the user selects several photos, track each upload separately so one failure does not erase the status of the others.

Handle expiration by requesting a new authorized upload opportunity when appropriate. Avoid unbounded retries, and prevent repeated completion requests from producing duplicate gallery records.

Keep verification on the server

Once storage reports success to the browser, ask the backend to finalize the upload. Verify the expected object, actual size, and required content checks before marking the item ready.

A client-side size check improves feedback but cannot enforce a trustworthy limit on its own. Design storage admission or processing controls for strict limits, and clean up abandoned objects. Direct uploads are a useful architectural choice when their operational benefits fit the application. They do not remove the server's responsibility for permissions, data integrity, or the photo lifecycle.