AI Artifacts & Generated Sites

No-Backend App

Definition

A no-backend app is a web app with no application server behind it. The visitor downloads HTML, CSS and JavaScript from a host, and from that moment every calculation, every screen change and every stored value happens on their own machine. There is no Node process, no PHP, no Rails, no shared database that the app can write to. Most things a chat assistant generates fall into this shape by default: a single-file web app that opens, runs and never phones home. It can still talk to the outside world through the browser's fetch API, but only to someone else's API, and only with credentials that the visitor could read if they looked.

Why It Matters

The distinction decides where your app can live and what it can promise. A no-backend app can be dropped onto any static host and served from an edge cache for nothing; an app with a backend needs a machine that stays awake, and that machine costs money and attention. The catch is on the data side. If two people open a no-backend to-do list at the same address, they each get their own private copy — nothing they type crosses between them. Teams discover this late, usually after showing a client a shared tracker that turned out not to be shared at all.

How It Works

The host answers a request for 'index.html' with a file and nothing else; no code of yours runs on the server. The browser parses that HTML, fetches whatever CSS and JavaScript it references, and executes the script. State goes into JavaScript variables for the session and into browser storage, usually localStorage, if it needs to survive a reload. Anything that genuinely needs a server — sending email, checking a password properly, writing to a shared table — has to be handed off to a third-party API called from the page, or dropped. 99helpers serves static files over HTTPS and runs no build and no server process, so a no-backend app is exactly what fits.

Real-World Example

A consultant asks Claude for an hourly-rate calculator, saves the result as 'rates.html' and publishes it at acme-rates.99helpers.site. It loads in under a second, works offline once cached, and remembers the last figures the user typed. What it cannot do is collect those figures centrally — so the consultant pairs it with a lead capture form on the same file, which does post somewhere, and stops pretending the calculator itself is a database.

Common Mistakes

  • Assuming data typed into the page is visible to you or to other visitors — it is not, it sits in that one browser until it is cleared
  • Putting an API key in the JavaScript because there is no server to hide it in — anyone can open view-source and read it
  • Describing a no-backend app as offline-capable without adding a service worker — a plain page still needs the network for its first load

Related Terms

Put a file online in seconds

Drop in a document, an image, a page or a whole static website and share the link — free, with no build step and no server to set up.

Host a file free →