CDN Service

The CDN Service is a Cloudflare Worker that serves Storybook builds from R2 storage, resolving the project and version from the request path and extracting only the bytes it needs from the stored ZIP.

Features

• Partial ZIP Extraction - Fetch only required bytes from ZIP archives
• Path Routing - One hostname, project and version resolved from the path
• Private Projects - Firebase-authenticated access, plus signed preview tokens for in-plugin previews
• Multi-Platform Support - Cloudflare Workers with R2 or Docker with filesystem
• Edge Caching - Global edge caching with Cloudflare KV
• SPA Fallbacks - Smart path resolution for single-page apps

Architecture

scry-cdn-service/
├── src/
│   ├── adapters/
│   │   ├── storage/              # R2 / filesystem backends
│   │   └── zip/                  # R2 range reader for unzipit
│   ├── services/
│   │   └── zip/                  # Central directory + extraction logic
│   ├── routes/                   # Route handlers
│   └── utils/                    # Shared helpers
├── cloudflare/                   # Cloudflare Worker entrypoint
└── docker/                       # Node server & Docker assets

How It Works

1. Request: https://view.scrymore.com/{projectId}/{versionId}/path/to/file.js
                    │
                    ▼
2. Parse the path for project and version, then check access
                    │
                    ▼
3. Load ZIP central directory from KV cache
   (or hydrate from R2 if cache miss)
                    │
                    ▼
4. Locate file entry in central directory
                    │
                    ▼
5. Fetch only the file's byte range from R2
                    │
                    ▼
6. Decompress if needed (stored vs deflate)
                    │
                    ▼
7. Return with correct MIME type and cache headers

URL Format

Storybooks are served from one hostname, with the project and version in the path:

https://view.scrymore.com/{projectId}/{versionId}/{path}

Examples, using the live demo Storybook:

• https://view.scrymore.com/U9m2H2yeC9wFiR4hlMta/demo-1786260947/
• https://view.scrymore.com/U9m2H2yeC9wFiR4hlMta/demo-1786260947/iframe.html
• https://view.scrymore.com/U9m2H2yeC9wFiR4hlMta/demo-1786260947/index.json

Performance

Metric	Value
Cache miss (first request)	~62ms
Cache hit	~31ms
Data transfer reduction	~50x vs full download
Edge locations	300+ globally

Storage Backends

R2 (Cloudflare)

Primary deployment path:

• Archives stored at static-sites/{project}.zip
• Partial extraction via range requests
• Metadata cached in KV

Filesystem (Docker)

For local development:

• Pre-extracted directories
• Mimics ZIP contents structure
• Hot reload support

Quick Start

Viewing Deployed Storybooks

# Access a deployed Storybook
curl https://view.scrymore.com/{projectId}/{versionId}/

# Access a specific file
curl https://view.scrymore.com/{projectId}/{versionId}/index.json

Self-Hosting

See Deployment for self-hosting instructions.

Next Steps

• Architecture - Technical deep dive
• Path Routing - URL routing details
• ZIP Extraction - Partial extraction explained
• Deployment - Deploy your own instance