platform Introduced in L4

Cloudflare Pages

Cloudflare's static + functions hosting. Built on Workers. Fast global CDN, fair pricing, generous free tier.

Mindmap

hover · click to navigate
this tech depends on / used by alternative Shipyard anchor
What it is

The plain-English version

Cloudflare Pages is a hosting platform for static sites and JAMstack apps with serverless functions. It's built on top of Cloudflare Workers (the edge runtime) and the same global CDN that serves much of the internet. Connect a git repo, push, deploy.

Why it exists

The problem it solves

Cloudflare's pitch is performance (every deploy is global out of the box), cost (the free tier is enormous, paid scaling is cheap), and integration (the same dashboard has DNS, R2 storage, D1 database, queues). For a non-Next-specific deploy, it's compelling.

What it competes with

Alternatives

AlternativeTypeWhen it wins
VercelplatformThe Next.js platform of record. Tightest integration with Next, expensive at scale, fastest dev iteration.
AWScloudThe largest cloud. Every primitive you'll ever need, plus a punishing bill if you're not careful.
Where it shows up in Shipyard

Deep links

Vocabulary

The words you'll hear

Workers
The edge runtime. V8 isolates instead of containers. Sub-millisecond cold start.
R2
Cloudflare's S3-compatible object storage. No egress fees.
D1
Cloudflare's SQLite-based edge database.
KV
Cloudflare's key-value store.
Wrangler
The CLI: wrangler pages deploy, wrangler dev.
Preview deployments
Each PR gets a unique URL automatically.
Prompting

Bad vs. good prompt for Cloudflare Pages

✕ Bad prompt
deploy this to cloudflare
✓ Good prompt
Deploy this Next.js (App Router) site to Cloudflare Pages. Use the @cloudflare/next-on-pages adapter. Set up the build command and output directory. Configure custom domain tasklane.example. List the env vars to set in the Pages dashboard. Note any Next features (image optimization, ISR) that need different handling than Vercel.

Why it works: Specifies the adapter (Next on Pages has its own quirks), the domain config, and prompts about feature parity gaps with Vercel — the actual gotchas.

Pitfalls

What bites real teams

⚠ Worker size limits

Free tier has tight limits on bundle size. Heavy npm dependencies blow them. Ship small, lazy-load on the client when needed.

⚠ Node API gaps

Workers don't have all Node APIs. fs, net, child_process aren't there. Some libraries assume them.

⚠ Cold-start surprises absent

Workers basically don't have cold starts — but if you're benchmarking against Vercel/Lambda thinking you'll see the same characteristic, you'll be confused.

References

Official docs only