Where the code runs
Where your code runs is a design decision, not an afterthought. For a small, AI-backed app growing out of a vibe-coded prototype, a serverless platform is a strong fit: you deploy code rather than manage a server, it costs nothing while idle, it scales up if the work takes off and it is cheap to start. Cloudflare Workers is one good example; Vercel or AWS Lambda are other common choices. It is not a universal answer though. These platforms run with execution-time and memory limits and a different runtime from a normal server, so heavy long-running compute or large in-memory state can fight the model. When that is your workload, a regular server or container may suit better.
Say it another way
Workers is a way to run your code on Cloudflare’s machines all over the world, instead of renting and managing a server yourself. You hand over the code, it runs close to your users, and you mostly pay for what you actually use.
Part of Designing the system.