coding
SWE-agent
Research coding agent from Princeton/Stanford that fixes GitHub issues; now superseded by mini-swe-agent.
The SWE-agent README says most development has moved to mini-swe-agent, which supersedes it.
what it is
SWE-agent is an open-source agent that lets a language model work on real software repositories, for example to fix GitHub issues. Its maintainers now focus on mini-swe-agent, which supersedes it.
who it's for
Researchers and developers experimenting with autonomous coding agents and benchmarks.
how to start
- 1Read the docs at swe-agent.com/latest.
- 2Consider starting with mini-swe-agent (github.com/SWE-agent/mini-swe-agent), the current focus.
- 3Install per the docs and add your model API key.
- 4Point it at a repository and an issue.
what you need
- Python
- An LLM API key
- Docker is typically used for sandboxes (see docs)
pricing
Free and open source
You pay your model provider for any API usage
checked Oct 10, 2026
key features
- Autonomous issue fixing
- Research-grade (paper on arXiv)
- Open source
- Superseded by mini-swe-agent
similar agents
- OpenHandssame job
- Julessame job
- Clineoverlapping features
- GitHub Copilotoverlapping features
- Claude Codeoverlapping features
- Devinsame job, different audience