Hardware

What each part needs.

CodeHero itself needs an ordinary computer. Anything more depends on your choices: where your applications run, whether AI models run on your own hardware, and which devices your apps are for.

The computer that runs CodeHero

The minimum is 1 CPU core, 4 GB RAM, 40 GB free disk and an ordinary internet connection; with 8 GB RAM it runs comfortably.

Ubuntu 24.04 — directly on Linux, or in a virtual machine on Windows or a Mac.

Installation guides

Where your applications run

While they are built: in a preview on that same computer.

In production, where you choose: on the CodeHero computer itself, with their own domains, SSL and backups — give it resources for them beyond the minimum — or on separate servers or hosting.

The gateway case study was installed on four separate servers.

Hosting and maintenance

Local AI models — only if you choose them

For the local models we have tested and recommend for this work, at least 96 GB of memory for the model. For example:

A computer with an AMD Ryzen AI Max+ 395 (128 GB) or Max+ PRO 495 (192 GB)
An NVIDIA DGX Spark, or another maker’s computer with the same chip
An Apple Mac Studio with at least 128 GB
A desktop with three NVIDIA RTX 5090, three AMD Radeon AI PRO R9700 or three Intel Arc Pro B70 cards (32 GB each)
Professional cards such as the NVIDIA RTX PRO 6000, and larger

Or AI services — no extra hardware

Connect an AI company’s service (by subscription or API key) and nothing more is needed on your side. That service receives the parts of the work it needs to do each step.

Where your data goes

Apps for Apple devices and Android

Apps for iPhone, iPad, Mac or Apple Watch are built on a Mac.

Apps for Android are built on a Windows, Linux or Mac computer.

The architecture underneath — the platform, the agents, the containers — is on the page for developers. For developers

Questions people ask

All the articles

Can small local AI models do real work?

Inside the factory, yes — more than on their own. Small, clear tickets, a checklist for finishing, tool help and a stronger model that checks the result let small models on your own hardware do much of the building. Read the article

Why use AI subscriptions instead of paying per use?

Because in our own work a subscription has cost up to about 40 times less than paying per use for the same work — around 30 times in some weeks, around 40 in others, depending on the models. The difference is always large. CodeHero runs the subscription AI tools (Claude Code, Codex, Grok Build) as workers in the factory. Read the article

Why does CodeHero host the applications too?

Because building and running belong together. The application runs where it was built, with its domain, secure connection, backups and recovery — and every later change returns to the same factory. No separate hosting panel to learn. Read the article