DeepYardDeepYard
S

SWE-agent

Princeton's autonomous agent for resolving GitHub issues and software engineering benchmarks

Open SourceFree

About

SWE-agent is an open-source autonomous software engineering agent developed by Princeton's NLP group. It pairs a language model with a custom Agent-Computer Interface (ACI) that enables structured interaction with codebases — reading files, editing code, running tests, and navigating repositories. Originally built to solve tasks from the SWE-bench benchmark, SWE-agent has become a standard research baseline and is actively used for automated bug fixing, patch generation, and GitHub issue resolution.

Details

Typecoding-agent, research, autonomous-agent
Deploymentlocal, cloud
Supported Modelsclaude, gpt-4, gemini, openrouter

Tags

open-sourceresearchswe-benchautonomous-agentbug-fixingcoding-agentacademic