cloudflarelisted

# Cloudflare ## Core Principle Workers are V8 isolates, not Node.js, the edge runtime is small, fast, cold; secrets via `wrangler secret put`, bindings over fetch. ## Iron Laws <EXTREMELY-IMPORTANT> - **Workers are V8 isolates, not Node.js.** No `fs`, no `child_process`, no `Buffer` (use `Uint8Array`). - **Edge runtime = small, fast, cold.** Keep bundles small. No kitchen sink imports. - **Secrets via `wrangler secret put`**, never in `wrangler.toml`. Never in code. - **Compatibility flags** for Node APIs (`nodejs_compat`, `nodejs_compat_v2`). - **Bindings over fetch.** Use D1/R2/KV bindings, not HTTP to own services. </EXTREMELY-IMPORTANT> ## When to Use Writing Workers; configuring `wrangler.toml`; bindings (KV, D1, R2, Queues); Pages; Workers AI; Tunnels; WAF; DNS; any Cloudflare infra. ## When NOT to Use Plain Node.js server (no CF); static without Workers; different platform. ## Workflow 1. Pick the service from the Core Services table. 2. Write the Worker with a typed `Env` binding contract. 3. Declare bindings in `wrangler.toml`; put secrets via `wrangler secret put`. 4. Iterate with `wrangler dev` (`--local` for fast iteration, `--remote` for actual Cloudflare bindings). ## Core Services | Service | Use | |----------------|-----------------------------------------------| | **Workers** | Compute at the edge (V8 isolate) | | **Pages** | Static + Workers Functions | | **KV** | Low-latency key-value (eventually consistent) | | **D1** | SQLite at the edge | | *