gpt-engineer
Pioneering CLI that turns natural-language specs into working codebases.
About
gpt-engineer is the original open-source code-generation experimentation platform by Anton Osika: specify software in natural language, watch the AI write and execute the code, then ask it to implement improvements. Launched in 2023, it defined the natural-language-to-codebase pattern that later tools and products iterated on; its own README notes that the opinionated managed-service evolution is lovable.dev, while the hackable-CLI lineage continued in tools like aider.
It installs via pip, supports Python 3.10 to 3.12, and is configured with a single API key. The canonical repository is AntonOsika/gpt-engineer (the gpt-engineer-org mirror exists but redirects). It is MIT licensed with over 55k GitHub stars, though development has slowed as the author focuses on commercial successors — treat it as a proven, influential reference implementation rather than an actively evolving product.
What it can do
- Natural-language to full codebase generation
- Iterative improvement prompts on generated code
- Code execution during generation
- Stable PyPI release plus dev install path
- Precursor lineage to lovable.dev
Setup
- pip install gpt-engineer (stable) or clone and poetry install (dev)
- Export OPENAI_API_KEY
- Run gpt-engineer in an empty folder with a natural-language prompt
Prerequisites
Python 3.10-3.12; an OpenAI API key.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via official GitHub repo (AntonOsika/gpt-engineer; gpt-engineer-org mirror redirects), batch277; GitHub 55,026 stars; development slowed as author focuses on commercial successors — influential reference implementation; Trending score 82/100 from researcher assessment. License: MIT.
★ 55,026 GitHub stars
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.