SW

SWE-agent

Language-model agent that resolves GitHub issues

Self-Hosted AI ★ 20k stars Medium setup MIT

SWE-agent lets a language model autonomously navigate a repository, edit files, and run tests to fix bugs or implement features. It is widely used as a research baseline for automated software engineering benchmarks.

Key features

  • Autonomous repository navigation
  • Custom agent-computer interface
  • Reproducible benchmark runs
  • Configurable LLM backends

Strengths

  • Released under the MIT license
  • First-class Docker support for quick deployment
  • Mature project with 20k GitHub stars
  • Written in Python

SWE-agent replaces

Compare SWE-agent

12 head-to-head comparisons.

Similar self-hosted ai apps