Request Flow
End-to-end request lifecycle: DNS → Cloudflare edge → dispatch worker → theme worker → Shopify origin (fallback). How routing decisions are made.
This is a living outline of planned documentation. Pages marked with a link are written; the rest are stubs to be filled in as we build.
How the system works under the hood.
Request Flow
End-to-end request lifecycle: DNS → Cloudflare edge → dispatch worker → theme worker → Shopify origin (fallback). How routing decisions are made.
Transpiler Pipeline
How Liquid templates become TypeScript render functions. IR stages, AST transformations, code generation, registry shim.
Cache Strategy
Per-section content-addressable caching. Cache key construction (settings hash + locale + market + deploy version). TTL strategy. Webhook-driven surgical invalidation.
Islands & Script Loading
Third-party app isolation, requestIdleCallback deferral, app block hydration. How scripts are categorized and loaded.
Data Pipeline
KV data model: products, collections, pages, blogs, menus. Storefront API queries, Admin API metafield enrichment. Fetch manifests.
Step-by-step instructions for common operations.
Client Onboarding
Setting up a new client: 4-char slug, Shopify custom app, D1/KV/R2 provisioning, .mise.local.toml, first build and seed.
Deployment
deploy:theme, deploy:dispatch, promotion pipeline, rollback, A/B testing. Manual steps when automation fails.
Local Development
pnpm dev, worktree conventions, port offsets, Miniflare state, cache clearing. Common gotchas.
Parity Workflow
pnpm parity:loop, section-by-section verification, Chrome extension visual diff, DOM comparison.
Adding a Liquid Filter or Tag
How to add support for a new Liquid construct: transpiler IR node, code generator, runtime helper, test fixture.
Webhook Invalidation
How product/collection/theme webhooks trigger surgical cache purges. Adding new webhook handlers.
Operational details. These pages contain sensitive configuration and are excluded from the public build.
Infrastructure Config
Cloudflare account IDs, API token locations, KV/D1/R2 namespace IDs, worker naming conventions. Per-client custom app credentials.
Client Runbooks
Per-client specifics: known template issues, coverage status, metafield manifests, contact details.
Incident Response
What to do when: worker goes down, cache is stale, Shopify changes an API, a deploy breaks production. Rollback procedures.
Manual Operations
Forced redeployment, manual webhook processing, D1 migrations, R2 cleanup, KV bulk operations.
Cost Tracking
Cloudflare usage per client, KV read/write volumes, Worker invocation counts, R2 storage.
As the platform scales, these will become important: