GPT Engineer
Conversational AI coding agent that builds entire codebases from natural-language specs with one prompt.
GPT Engineer
GPT Engineer turns a plain-English project description into a working codebase. It prompts you for clarifications, generates all necessary files, and can iteratively improve the code through a feedback loop — all from the command line.
Key Features
- Full-project generation — produces complete, runnable codebases (not just snippets) from a single spec file
- Clarification loop — asks targeted questions before generating to reduce ambiguity
- Improve mode — run it on an existing repo to add features or fix bugs
- Benchmarked — evaluated on HumanEval and SWE-bench style tasks
- Provider agnostic — supports OpenAI, Anthropic, Azure OpenAI, and local models via LM Studio / Ollama
- Lightweight — no heavy orchestration framework; pure Python with minimal dependencies
Quick Start
pip install gpt-engineer
gpt-engineer projects/my-new-app
Create projects/my-new-app/prompt with:
Build a FastAPI web app with a /health endpoint and SQLite persistence for user notes.
Install via ai-supply
npx ai-supply add gpt-engineer-code-generator
Curated mirror of the open-source GPT Engineer (MIT). Get it from the source.
Compromise signals — malicious or tampered code (leaked secrets, backdoors, a dropped executable) — reduce the score, and known dependency CVEs carry a bounded penalty (they warrant review but never QUARANTINE — update the dependency to clear). Other dangerous-by-capability traits are risk surface, expected for some capabilities. Every finding is mapped to its OWASP control below.
Findings mapped to the OWASP Top 10 for LLM Applications (2025) and the OWASP Machine Learning Security Top 10. Expand any flagged control for the exact findings — compromise reduces the score; expected/risk-surface do not, except a known CVE, which carries a small bounded penalty (high/critical → Review).
The same gate an agent runs before installing (POST /api/v1/trust/gpt-engineer-code-generator/check). Click a policy:
Consume GPT Engineer programmatically. Authenticate with an API key or session — see Authorize an agent.
# Agents: CHECK BEFORE YOU INSTALL (no auth) — score, grade, level, capability manifest
curl https://ai-supply.store/api/v1/trust/gpt-engineer-code-generator
# Gate against your org policy (returns { pass, violations })
curl -X POST https://ai-supply.store/api/v1/trust/gpt-engineer-code-generator/check \
-H "Content-Type: application/json" \
-d '{"minGrade":"B","denyPermissions":["shell"],"denyUnknownEgress":true}'
# CLI
npx ai-supply add gpt-engineer-code-generator
# REST (install → download)
curl -X POST https://ai-supply.store/api/v1/listings/gpt-engineer-code-generator/install \
-H "Authorization: Bearer $AIM_KEY"
# MCP tool
install_listing({ "slug": "gpt-engineer-code-generator" })OpenAPI spec →Curated mirror — latest upstream source. See the repository for tagged releases.