Ask an AI assistant to scaffold a small app like a habit tracker or counter from scratch.
Collaborate live with an AI on one running app, watching its edits appear instantly.
Package the same app for web, iOS, Android, desktop, or a mini-program without rewriting it.
Distribute a signed app bundle that verifies it has not been tampered with before running.
| qorm/qorm | aeneasr/github-trends | bxcodec/go-simple-flatbuffer | |
|---|---|---|---|
| Stars | 4 | 4 | 4 |
| Language | Go | Go | Go |
| Last pushed | — | 2020-12-18 | 2017-06-09 |
| Maintenance | — | Dormant | Dormant |
| Setup difficulty | — | easy | easy |
| Complexity | — | 2/5 | 2/5 |
| Audience | vibe coder | pm founder | developer |
Figures from each repo's GitHub metadata at analysis time.
QORM is a tool for building small apps, like a counter or a to-do list, where an AI coding assistant such as Claude or Cursor does most of the work of scaffolding, editing, and testing the app for you. The idea is that you and the AI work on the same running app at the same time: when the AI makes a change, you see an on-screen notice and the update appears right away, and when you click something in the app, the AI can see that too. Under the hood, apps are described as small JSON files rather than in a specific programming language, which is what lets the same app run on the web, iOS, Android, desktop computers, and even mini-programs, all built from one description. The project is written in Go, and by default it compiles to plain Go with no extra runtime dependencies, so a built app is a single small file. There is also an optional native desktop window mode for people who want their app to feel like a normal installed program instead of running in a browser tab. QORM connects to AI assistants using the Model Context Protocol (MCP), a way for AI tools to call functions and see results. Through this connection, the AI can inspect the app's current state, run actions like button clicks, preview a proposed change before applying it, and check that a change actually worked. Changes can also be packaged into signed bundles, so the app verifies they have not been tampered with before running them. This project would suit someone who wants to experiment with letting an AI assistant build and modify a working app live, rather than only writing code that a human reviews afterward. It assumes some comfort with running command line tools like go install and go run, and with the general concept of connecting an AI assistant to a project. People who have never used a terminal or an AI coding tool before will likely find the setup unfamiliar without extra guidance.
A tool that lets an AI assistant build and live-edit small cross-platform apps with you, described as simple JSON instead of code.
Mainly Go. The stack also includes Go, MCP, WebView.
Mainly vibe coder.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.