Save AI Videos Locally: Folder Access, Collection Receipts and Retention Notices

Save from the retention notice

This retention policy applies to registered results from new backend tasks, not automatically to every asset. Do not assume ordinary uploads or direct URL results are covered. When the site has local media enabled and you have items needing attention, a notice appears after sign-in. Find the project and asset, click Save to this device, then select a folder and allow read/write access in a Chromium browser that supports the File System Access API.

You must initiate authorization with a click. Background work does not open a folder picker on its own; lost permission needs your confirmation again. If the browser or device lacks this API, do not expect automatic collection. Use Download in the notice to save a file instead.

After writing the files successfully, collection sends a receipt to the server. It identifies the device, object version and file checksums to confirm collection; it is not a permanent cloud backup.

A download and a collection receipt are different

Download uses the browser's download path. Save to this device uses folder collection and sends a receipt. An ordinary download is not proof that receipt-backed collection completed. Open the file, check that it plays fully and place it in your project media folder.

Local files depend on this device and the folder you chose. Another computer will not automatically find them, and clearing browser data or losing folder permission can interrupt access. Keep a separate backup of important footage. Playback on a canvas does not replace checking the file on disk.

The 15-day period, reminders and one extension

The default retention period for unclaimed media is 15 days. Only when an administrator enables unclaimed cleanup does the service schedule in-app and email notices three days before expiry (T-3) and one day before expiry (T-1). Email delivery also depends on notification settings and the mail service. Save the files when notified; a reminder is not a backup.

For an unclaimed item that has not been extended, click Extend 15 days in the notice to add 15 days once. Extensions do not stack. The button is unavailable for collected items, previously extended items or ineligible states. Use the expiry date shown for that asset.

Both confirmedCleanupEnabled and unclaimedCleanupEnabled are off by default, so day 15 does not automatically delete the file. If an administrator enables cleanup, unclaimed media must also be past its expiry, have successful T-3 and T-1 in-app notices and attempted email notices, and pass the remaining receipt and protection checks. This is not a promise of long-term server storage.

Make collection part of finishing a shot

After a shot completes, review it, save it and verify the local file. Organize footage by project, episode and shot number, and keep the version you chose. Avoid moving files referenced by an editing project.

Notification options in the notice control preferences, not retention periods or cleanup policy. If you miss a reminder, inspect the asset's current state. Local collection cannot restore a server copy that is no longer available.

Connect your own API channels and third-party relays

In model-channel settings, enter the Base URL and API key, then configure the protocol, model IDs and required headers from your provider's documentation. Check the image and video modes you need before generating. Paths, resolutions and first/last-frame inputs depend on the connected channel. An adaptation request for an unsupported channel is not a compatibility guarantee.

BYOK supports a custom API endpoint, model key and headers. For OneAPI, NewAPI or another relay, select the OpenAI, Gemini or Claude protocol that matches its actual API, then check model IDs and image/video endpoint support. Not every relay or private protocol works automatically. BYOK model calls have no platform model markup or platform generation-credit deduction; workspace access and provider usage are still billed separately.

Connect only trusted providers and use HTTPS in production. Custom requests are forwarded through the authenticated backend; required prompts, reference assets and credentials take part in the request. Never put API keys, passwords or secret-bearing headers in public canvases, screenshots or support requests.