EdgeCompute

Code running at CDN edge locations, close to the user, instead of a centralized region — lower latency, different constraints (no persistent filesystem, limited runtime).

Why we need this / what value this brings

Running code physically closer to the user cuts network latency that no amount of backend optimization can fix.

When to use this

Latency-sensitive logic that doesn't need heavy compute or a large runtime (auth checks, redirects, simple transforms, the BYOK chat's API route).

How to use or implement this

Keep edge functions lightweight and stateless; be aware of the runtime's constraints (no Node-specific APIs, limited CPU time on Cloudflare Workers).

Research questions

• What Cloudflare Workers-specific constraints (CPU time limits, no Node APIs) actually shape what can run at the edge?

---

Empty folder — drop notes, links, and findings here as you research.