Bundle up an entire directory, and distribute it across Cloudflare's range of products, including embedded in the Worker, KV, and R2!

This uploads the necessary files & outputs a functional Cloudflare Worker that handles routing to the uploaded assets. You can use use Hono, the web framework/router, or Vanilla, handling the fetch event ourselves with no extra packages.

• If you don't have assets over KV/Cloudflare Page's limit of 25 MB, you probably are better off just using Cloudflare Pages, or Worker Sites.
• If you do have assets over KV's limit, or want to exclusively use R2, then this allows you to upload a static directory anywhere.
• If you want to spread your assets out by size / directory / extension across KV and R2, even multiple buckets/namespaces, without having to manage it yourself.
• This also lets you (optionally) embed assets inside of the worker itself, base64 encoded. The ultimate unnecessary speed boost.
• Lastly, this optionally supports a "Manifest", keeping track of all old uploads, and removing unused files. We use this for avoiding uploading the same file twice, and also to keep assets around for a few versions (MaxManifestCount, default 3), in case your worker deployment goes wrong/worker propagation is slow and your old scripts are still around.

• Download the executable (or build it), and run it once. It will generate a bundleconfig.json for you to configure.
• Set up the storages / etc you want to use. You can have multiple of the same storage if you want different settings. See the configurations below, or the full configuration example below.
• Run the script with bundle (i.e ./CloudflareSuperSites bundle), and it will output a (hopefully) usable worker. Deploy the worker via Wrangler (wrangler publish)! If you want to set up CI/CD (not that I would recommend it, this may upload junk/a broken script): check here!
• You can use the clean up command (./CLoudflareSuperSites cleanup) with a manifest storage provider configured to clean up old files & manifests.

Notes:

• These must all match for the asset to be included. For example, if you specify to include the file path /cookies/, and only HTML file extensions are allowed, it must match both
• Include Paths based on the case-insentive Start of the relative path. For example, if your website is example.com, and you have a path at example.com/cookies/, IncludePaths /cookies/ would match any assets under that path
• File Extensions are also case-insensitive

There are no unique embedded storage configuration options, just the base storage configuration options.

Why the S3 API Keys? We use Minio under the hood, there's no offical CF API for R2, and the unofficial one has an upload limit of ~100 MB (as does the entire API!). There are bits of code left for using the unofficial API, if you see them in the source.

Todo:

• Tests
• Maybe Tests with unstable_dev/the worker itself?
• Support for serving old assets (could force uploading all to a persistent data store, and re-creating routes based on the manifest)
• A Non-Full Worker mode, that outputs code that appends to your existing worker / otherwise lets you hook into it and use SuperSites alongside a normal worker (like you can use Cloudflare Worker Sites alongside a worker)

Full Configuration Example:

{ "BundleDirectory":"../public", # Relative or Exact Directory "OutputLocation":"worker/worker.js", # Relative or Exact File Path "Router":"Vanilla", # Vanilla or Hono "StorageConfigurations":[ # Repeat mutiple of the same type or exact same namespaceid/bucket name if you want. { "Type":"Embedded", "AllowedFileExtensions":[ "html", "css", "js" ], "FileSizeLimit":200000 }, { "Type":"KV", "FileSizeLimit":10000000, "ApiToken":"...", # Use Environment Variables if your environment supports it. KV needs KV Edit Permissions "AccountId":"...", "NamespaceId":"...", "BindingName":"KV", }, { "Type":"R2", "AccessKey":"...", # R2 S3 API Keys "SecretKey":"...", "AccountId":"...", "BucketName":"main-website", "BindingName":"R2", "CacheSeconds": 3600 } ], "MaxManifestCount":3, "ManifestStorageConfiguration":{ # Not Required, if you don't care about leaving files for old deploys behind/leaving unused files in storage "Type":"R2", "BucketName":"main-website-manifest", "AccessKey":"...", "SecretKey":"...", "AccountId":"...", "BindingName":"R2", } }

Example Site using Super Sites: https://tylerobrien.dev