A Cloudflare Worker for serving and browsing content in a private Backblaze B2 bucket.

• Serve the content in your B2 bucket for free
• Browse the content in your B2 bucket

• Backblaze B2 is cheap file storage
• Downloading from Backblaze B2 is normally $0.01 per GB
• Downloading from Backblaze B2 through Cloudflare's CDN is free
• Private buckets prevent people from downloading from your bucket directly

See Cloudflare's documentation for more information.

Copy wrangler.toml.example to wrangler.toml and fill in your own information into the variables. Refer to Cloudflare's documentation for more information.

• routes The domains and subdomains this Worker will listen on. This needs to contain at least the domains specified in the MAIN_DOMAIN and DIR_DOMAIN vars below.
• vars The following vars are required for the Worker to function correctly.
  • B2ACCESS Application key ID from Backblaze B2.
  • B2BUCKET The name of the B2 bucket to use.
  • MAIN_DOMAIN The domain from which files will be served. If a request comes in for any other domain (except DIR_DOMAIN), it will be redirected to this domain.
  • DIR_DOMAIN The domain to use for directory listings.
• kv_namespaces Create a Workers KV namespace and bind it to this app with the name B2CDN
• triggers B2 authorization tokens only last 24 hours. When triggered by a cron job, this Worker gets a new authorization token from the B2 API and stores it in Workers KV.

Secrets are like vars but... secret. This Worker only needs one secret named B2SECRET which should contain the applicationKey (the secret part) of your Backblaze B2 application key.

You can use Cloudflare Access to require a login for certain subdomains. Let's say you want to publicly share the files in your B2 Bucket, but not allow others to list the files in it, you can set up a separate DIR_DOMAIN and place Cloudflare Access in front of it. There's a lot of supported identity providers. You can make it so only a certain IP can access the directory listing, or only particular Google Accounts.

• Implement pagination for enormous folders
• Implement more efficient calls to b2_list_file_names

MIT