AI Agents
Agent-Reach
Give your AI agent eyes on the whole internet: Twitter, Reddit, YouTube, GitHub and more, one CLI
★ 94.0k starsMIT
View guide →Give it a GitHub issue and it tries to fix the code for you - the original SWE-bench agent
SWE-agent is an open-source ai agents project by SWE-agent (SWE-agent/SWE-agent on GitHub). It has 20,506 GitHub stars and 2,247 forks, is released under the MIT license, and was last updated on GitHub on 6 October 2026.
Choose a real GitHub issue to attempt, or a SWE-bench split to benchmark against.
Clone the repository, or try it in your browser with the GitHub Codespaces link in the README.
Install from source per the docs and configure the LM you want to drive the agent.
Follow the docs' hello-world command to run the agent on one issue and review the patch it produces.
Reads a GitHub issue, navigates the repo, edits code and produces a candidate fix.
Works with the language model you configure rather than locking you to one provider.
Runs and benchmarks the agent across SWE-bench tasks for reproducible research results.
A dedicated mode adapts SWE-agent for offensive cybersecurity and capture-the-flag style challenges.
Try it in the browser via GitHub Codespaces without a local install.
The agent-computer interface design comes from the peer-reviewed SWE-agent paper (NeurIPS 2024).
Get the source code from GitHub and enter the project directory.
git clone https://github.com/SWE-agent/SWE-agent.git
cd SWE-agentThe README links to the docs' source installation guide.
pip install -e . # see README/docs installation for the exact commandThe documentation's hello-world page runs the agent on a sample issue from the command line.
# Follow the 'Hello world from the command line' guide linked in the READMEIt is maintained by the SWE-agent team (John Yang, Carlos Jimenez and collaborators) on GitHub.
Yes - SWE-agent is MIT licensed and free to use.
You supply the language model: most setups use a paid provider API key (or a hosted/local model you configure), and long issue-solving runs can consume significant tokens.
Compared with commercial coding agents, SWE-agent is a research-first, fully configurable harness - ideal for benchmarking and modification, less polished as an everyday IDE assistant.
Try the Codespaces link in the README, or install from source and run the docs' hello-world example on a sample issue.
Give your AI agent eyes on the whole internet: Twitter, Reddit, YouTube, GitHub and more, one CLI
The benchmark that grades AI on real GitHub issues.
Give your AI coding tool expert Swift Concurrency guidance.
Exa's MCP server for web search and crawling — give your AI agent a live, high-quality web index.
Monid — OpenRouter for agent tools: one base URL and one key give an agent 2,000+ tools across 72+ providers
Self-hosted AI pentesting agent that finds flaws and opens PRs to fix them.
The most-starred other AI Agents projects in the LibreAI catalogue — the realistic shortlist if SWE-agent is not the right fit.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Free and open — grab the source, star the repo, and start creating.