langchain-ai/open-swe is an open-source framework for building and deploying internal asynchronous coding agents that integrate with organization-specific tools and codebases. It provides a customizable architecture featuring isolated cloud sandboxes, multi-platform invocation, subagent orchestration, and automatic pull request creation.
Project overview
The project offers a deliberately curated toolset of approximately 15 tools combined with multi-platform invocation across Slack, Linear, and GitHub, addressing the organizational need for coding agents that operate within existing internal workflows.
Project type
AI Agent
Use cases
Coding & Development · Automation
Deployment
Refer to project documentation
License
MIT
Best for
Enterprise teams and developers who need to build and deploy their own internal asynchronous coding agents integrated with existing tools such as Slack, Linear, and GitHub.
AI engineers seeking a customizable open-source coding agent architecture with subagent orchestration and automatic PR creation.
Key capabilities
Commits changes and opens a draft PR linked back to the source ticket when the agent completes its task.
Getting started
Setting up a local development environment for the backend and dashboard is required. The overall setup difficulty is rated hard because it involves local development setup of both the backend and the dashboard and requires GitHub App creation.
A GitHub App must be created as part of the initial setup process.
A configured Large Language Model (for example, OpenAI) is required. Paid services from model or sandbox providers are optional.
Evidence and sources
README: Open-source framework for building your org's internal coding agent.
README: [Customization Guide](docs/CUSTOMIZATION.md) — swap the sandbox, model, tools, triggers, system prompt, and middleware for your org
README: The Deep Agents framework natively supports spawning child agents via the `task` tool. The main agent can fan out independent subtasks to isolated subagents
README: Every task runs in its own **isolated cloud sandbox** — a remote Linux environment with full shell access. The repo is cloned in, the agent gets full permissions, and the blast ra…
README: - **Subagent support** — the agent can spawn child agents for parallel subtasks
AI Search
Find projects, verify facts, compare options, or turn a complex need into an actionable plan
Try a searchA click only fills the search box; you stay in control