VerlumoAIAI coding agents that write, test, and review code across your whole codebase.
View detailsSWE-agent
An open-source agent framework that automatically fixes GitHub issues and repairs bugs in real codebases.
SWE-agent is an open-source agent framework that automatically fixes GitHub issues and repairs bugs in real codebases. Given an issue description and a repository, it searches the code, reproduces the problem, writes a fix, and runs the tests to check its work, all through a loop of actions in the terminal.
Key Features
- Automatically fixes GitHub issues in real codebases
- Reproduces bugs and runs tests to verify fixes
- Open-source framework you can extend and study
- Works with many models through your own API keys
- Terminal-based agent loop with full transparency
- Basis for the well-known SWE-bench benchmark
Pricing
| Plan | Price | Details |
|---|---|---|
| Free | $0 | Free and open source; API costs depend on the model you choose. |
Prices as listed in 2026 research. Always check the official site for the latest price — plans change. Free plan: Yes — it has a free plan.
Best for: researchers and developers who want an open-source agent framework for automatically fixing GitHub issues.
Popular with: Developers, Students
Popular with: Developers, Students
How it works
SWE-agent became well known through the SWE-bench benchmark, where it was one of the first systems to solve real-world GitHub issues automatically. It is designed as a framework: researchers and developers can change its tools, prompts, and models to test new ideas about how coding agents should work.
The project is free and open source, and you pay only for the model API usage you choose. It is best for researchers and technically confident developers who want to experiment with autonomous bug fixing rather than a ready-made product.
Pros and Cons
✓ Pros
- Free and open source
- Proven approach on real GitHub issues
- Fully transparent and extensible
- Good for research into coding agents
✗ Cons
- A research project, not a polished product
- Needs setup and your own model API keys
- Success rate varies a lot by codebase and issue
Frequently Asked Questions
Is SWE-agent free?
Yes — SWE-agent has a free plan. Free open source; you pay your own model API costs
What is SWE-agent best for?
researchers and developers who want an open-source agent framework for automatically fixing GitHub issues.
What kind of tool is SWE-agent?
SWE-agent is a Coding Assistants tool on VerlumoAI.
Related Tools
AI code reviewer that refactors Python and reviews every pull request.
View detailsVisual full-stack builder that turns designs into real, ownable code.
View detailsLast reviewed: November 2026