Puter.js - The Backend for AI-Generated Apps

Puter.js gives your AI-generated app a safe, scalable, production-ready backend. It is a revolutionary serverless, keyless JavaScript SDK that brings powerful cloud and AI capabilities directly to the frontend code.

Puter.js is powered by Puter, the open-source "Internet Computer".

AI can build complete, working apps with Puter.js in a single shot, because there is nothing to configure. There are no API keys to obtain, no servers to provision, and no SDK setup, so the AI never has to pause to ask you for credentials or generate backend code that you then have to deploy.

Puter.js is keyless and serverless by design. Auth, cloud storage, databases, and 500+ AI models are all available from a single <script> tag, and its small, consistent API is easy for language models to use correctly. Puter even publishes an llms.txt so AI tools can read the full API in one request.

And with the User-Pays model, AI-generated apps scale to any number of users without you paying for infrastructure.

No. Puter.js is a general-purpose backend SDK, and it is a great fit whether your code is written by an AI or entirely by hand. Any web app that needs auth, cloud storage, a database, or AI models can use it.

The same things that make Puter.js easy for AI to use, no API keys, no servers, and a small consistent API, make it just as pleasant for developers writing every line themselves.

All of them! Puter.js works with every AI coding assistant, agent and vibe-coding platform, including Claude Code, Codex, Cursor, GitHub Copilot, Google AI Studio, v0, Bolt, Lovable, and others.

There is no special integration or plugin to install. When you ask the AI to build an app, simply tell it to use Puter.js:

use Puter.js (more info if needed: https://docs.puter.com/llms.txt)

Because Puter.js needs no API keys or setup and is plain frontend JavaScript, any agent that can write HTML and JavaScript can produce a working app.

Whether one user or a million, apps built on Puter cost the developer nothing to run! Puter introduces a groundbreaking approach to infrastructure costs, called the "User-Pays" model, in which users pay for their own usage. Here is how it works:

1. Every user, once authenticated, gets their own allocated resources (storage, database, AI credits, etc.) as part of their account.
2. Apps tap into the user's pre-allocated resources, rather than requiring separate infrastructure.
3. If a user exceeds their resources, they pay Puter directly for the overage.
4. Developers pay $0 for infrastructure, regardless of user count.

This model means developers don't need to provision or pay for separate infrastructure for their apps, regardless of how many users they have.