open swe
Basic Information
Open SWE is an open-source, cloud-hosted asynchronous coding agent built with LangGraph that autonomously understands and modifies codebases. It is designed to take a coding task from initial planning through implementation by generating a multi-step plan, executing code changes across a repository, and opening pull requests. The project provides a web UI for creating and managing tasks and supports starting tasks from GitHub by labeling issues. The demo requires users to supply their own LLM API keys. Documentation and announcement materials are linked from the README for setup and usage. The repo targets automation of repository-wide programming workflows rather than being a library for local code editing.
Links
Stars
4191
Github Repository
Categorization
App Details
Features
Open SWE emphasizes a distinct planning phase that produces editable plans before execution. It supports human-in-the-loop interactions, allowing messages and feedback during planning and execution. Tasks run in a cloud sandbox and can execute in parallel without local resource limits. The agent performs end-to-end task management by creating GitHub issues and opening pull requests that close issues on completion. Users can operate Open SWE via a web UI or trigger tasks from GitHub labels such as open-swe and open-swe-auto, with specialized higher-performance labels available for more complex tasks.
Use Cases
Open SWE helps teams and individuals automate complex repository changes by producing actionable implementation plans and carrying out code edits across projects. The planning step gives reviewers the opportunity to accept, edit, or reject proposed changes, reducing risk. Real-time human-in-the-loop feedback enables iterative correction while the agent runs. Cloud sandboxing and parallel execution let users scale many tasks simultaneously without local compute constraints. Built-in GitHub integration automates issue tracking and pull request creation, streamlining workflow from task inception to merge. The web UI and GitHub-triggered workflows lower the barrier to adopting automated, repository-level code change automation.