Using your own bucket

A project can save new uploads to a bucket you own on Cloudflare R2 or Amazon S3. You can copy everything it already has into that bucket, then switch the project so it reads from there.

Why use your own bucket

A studio, a financier or a broadcaster may require the footage to live in an account they can audit. With your own bucket, the files are written there from the start.

Only the project owner can set this up, and it needs a Pro or Max plan.

Connect a bucket

The Custom Storage form
  1. Pick the provider

    In Custom Storage on the Data tab, set Provider to Cloudflare R2 or AWS S3.

  2. Say where the bucket is

    For R2, enter your Account ID. For S3, enter the Region, such as us-east-1. The endpoint under the field fills itself in.

  3. Enter the bucket and keys

    Fill in Bucket Name, Access Key ID and Secret Access Key. The key needs to read, write and delete in the bucket.

  4. Test and save

    Press Test & Save. Before it saves, the app writes a small test file, reads it back and deletes it. When that works, the box reads Connected.

If the test fails, the message says why: Bucket not found, an access denied message, or a connection error.

After it is connected

A connected R2 bucket
  1. The box shows the provider, the bucket and a masked key ID. The secret key is never shown again.
  2. Edit opens the form again. Enter both keys again to save a change. You can’t point the project at a different bucket while files are still in this one.
  3. Remove disconnects the bucket. New uploads go back to our storage, and files already in your bucket still open while its keys work.

The keys are stored encrypted and used only by our servers. Crew members never see them, not even those who can upload media.

Connecting or removing a bucket only changes where new uploads go. Files uploaded earlier stay where they are until you copy them.

Copy your media and switch

Under a connected bucket, Copy your media puts a copy of everything the project has stored into your bucket: takes, photos, generated images and documents. The copy runs on its own, so you can leave the page.

  1. Copy

    Press Copy media to my bucket. The box counts the files as it finds them, then as it copies them. Files already in your bucket are left alone, so running it again only copies what is missing.

  2. Check the result

    When it finishes, the box says how many files were copied. If any failed, it lists them. Fix the cause, often the key’s permissions, and press Copy again.

  3. Switch

    After a copy with no failures, press Switch this project to my bucket, then Switch. From then on the project reads and saves everything in your bucket. A last short copy picks up anything made since the first one.

Our copy of the files stays where it was. Keep the bucket and its keys working after you switch: if they stop, the project’s media stops loading.

Still stuck?Ask in the Discord