GPT-Engineer: generate an entire software project from a single sentence
In one sentence Anton Osika publishes GPT-Engineer on GitHub: describe what you want in natural language, the agent asks clarifying questions, then writes all the files and runs them. 50k stars in one week. First viral implementation of the 'one-shot project generator' concept.
Anton Osika, a Swedish developer, publishes a very simple tool on GitHub: write in a text file what you want to build ("a REST server that manages a shopping list with JWT authentication"), run a command, and the agent asks a few clarifying questions, then generates the entire project structure: folders, files, working code.
It is not perfect. The generated code often has bugs, dependencies need fixing, and for complex projects it needs guidance. But for the first time a tool makes accessible the idea of "speaking software into existence" without writing a single line of code manually.
In seven days it collects 50,000 GitHub stars, a growth rate never seen for an individual project. It precedes Devin by nearly a year, and lays the groundwork for all the "AI scaffolding" tools that follow: Smol Developer, GPT-Pilot, Cursor Agent, Claude Code.
The simplicity is the point: one spec file, one LLM, one writing loop. No complicated interface.
Companies
—
Tools
—
Tags
Sources