SWE-agent

What it does

SWE-agent lets a language model of your choice, such as GPT-4o or Claude Sonnet 4, autonomously use tools to fix issues in real GitHub repositories, find cybersecurity vulnerabilities, or perform custom tasks. Its behavior is governed by a single YAML file, and it is documented as simple and hackable by design.

Typical use is pointing a model at a GitHub issue and letting it attempt a fix, starting with the Hello world guide. The project leaves maximal agency to the language model and is built and maintained by researchers from Princeton University and Stanford University. Multimodal support allows processing images from GitHub issues with vision-capable models.

It suits researchers and developers who want to experiment with autonomous coding agents and can work with a model provider and a local Python installation. The project states that SWE-agent is now in maintenance-only mode and has been superseded by mini-swe-agent, which it describes as simpler while matching SWE-agent's performance, so new users are directed there.

What it can do

• Uses a chosen language model to autonomously operate tools against real GitHub repositories
• Solves GitHub issues in repositories through an agent workflow
• Finds cybersecurity vulnerabilities as a supported task
• Performs custom tasks defined by the user
• Processes images from GitHub issues with vision-capable models (multimodal support, July 9 release)
• Configurable behavior governed by a single YAML file

Works with

• GitHub

Before you choose

• The project is in maintenance-only mode and has been superseded by mini-swe-agent; consider mini-swe-agent for new work.
• Results depend on the language model you select, and model API access and costs must be arranged separately.
• Designed for research and hackability, so it is oriented toward experimentation rather than a packaged end-user product.

Listed under

• Coding
• Coding agents
• Bug fixing
• Coding agent
• Cybersecurity
• Developers
• GitHub
• GitHub issues

Get started

Install SWE-agent locally from source using pip; the docs describe this as the default option.

1. Install SWE-agent locally from source using pip, the default option described in the installation guide.
2. Alternatively, run it in a browser-based VSCode environment for a quick first look.
3. Follow the Hello world guide to solve a GitHub issue with SWE-agent.

Model API or compute costs for the chosen language model are a separate cost the buyer must verify.

We couldn’t find published prices for SWE-agent.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Getting Started - SWE-agent documentation ↗swe-agent.com
• Setting up SWE-agent - SWE-agent documentation ↗swe-agent.com

Company website

swe-agent.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Human review

Checked by a person

knowledge_base

manual_review

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.