GPT Engineer
The viral 2023 open-source CLI that generated whole codebases from one prompt.
Rebranded
product rebranded to Lovable late 2024
About
GPT Engineer, released by Anton Osika in June 2023, was the project that proved AI could generate an entire codebase from a single natural-language prompt. You described the app, it asked clarifying questions, then wrote the full file set in one batch. It passed 50,000 GitHub stars within months, making it the most-starred autonomous-coding project of its era.
The open-source repo is now archived and read-only; the team's energy moved to the hosted product gptengineer.app, which was rebranded to Lovable in late 2024. Included here as a historical landmark: the package still installs from PyPI and works as a hackable codegen sandbox, but new projects should look at maintained successors.
What it can do
- One-shot full codebase generation from a natural-language prompt
- Interactive clarification step before generation
- Improvement loop over existing project folders
- Benchmark binary for APPS and MBPP evaluation
Setup
Install with pip install gpt-engineer; archived repo, so prefer maintained successors for new projects.
Prerequisites
Python 3.10 to 3.12 and an API key for a supported model provider.
Use cases
Works with
Trust score — 6/10
Scored 6/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 86; Trending score 58/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.