OpenHands
Autonomous AI software engineer that codes, browses docs, and submits PRs
View on GitHubWhat it solves
The gap between 'AI can help me code' and 'AI can act as a coding agent': OpenHands closes it with a full execution sandbox where the agent takes real actions, not just suggestions.
Overview
OpenHands (formerly OpenDevin) is a platform for AI agents that act like junior software engineers: they read a GitHub issue, write code, run tests in a sandboxed Docker environment, and submit a pull request. The agent sees a full terminal, browser, and file system. Published at ICLR 2025. A v1 modular SDK lets you compose the planning, execution, and memory layers independently.
Key facts
- Language
- Python
- License
- MIT
- Maturity
- Fast-moving
- Maintainer
- OpenHands
- Reviewed
- 2026-06-11
Where to start
The agenthub README explains the planning agent types and when to use each, CodeAct is the default; read this before configuring a custom agent.
openhands/agenthub/README.mdTechnologies
More in this domain
LangGraph
Stateful multi-agent orchestration for LLM applications
The state management problem in multi-agent systems: how you build agents that pass context across turns without losing the execution thread.
AutoGen
Multi-agent conversation framework from Microsoft Research
How to coordinate multiple specialized agents toward a shared goal without writing low-level orchestration logic by hand.
CrewAI
Role-based multi-agent orchestration for collaborative task execution
The boilerplate problem in multi-agent setup: defining roles, delegation rules, and task routing without writing a custom orchestration graph.