ScriptsApr 6, 2026·2 min read

OpenHands — Open-Source AI Software Developer

Autonomous AI software developer that can write code, run commands, browse the web, and interact with APIs. Formerly OpenDevin. Resolves 53% of SWE-bench Verified tasks. 45,000+ stars.

AG
Agent Toolkit · Community
Quick Use

Use it first, then decide how deep to go

This block should tell both the user and the agent what to copy, install, and apply first.

# Docker (recommended)
docker pull docker.all-hands.dev/all-hands-ai/runtime:0.20

docker run -it --rm \
  -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.20 \
  -v /var/run/docker.sock:/var/run/docker.sock \
  -p 3000:3000 \
  docker.all-hands.dev/all-hands-ai/openhands:0.20

Open http://localhost:3000. Set your LLM API key and start assigning tasks.

Intro

OpenHands (formerly OpenDevin) is an open-source autonomous AI software developer that writes code, runs commands, browses the web, and interacts with APIs with 45,000+ GitHub stars. It achieves 53% on SWE-bench Verified — the highest score among open-source agents — meaning it can resolve over half of real-world GitHub issues autonomously. It runs in a sandboxed Docker environment for safety. Best for teams who want to delegate real development tasks to AI. Works with: Claude, GPT-4, any OpenAI-compatible model. Setup time: under 5 minutes with Docker.

Capabilities

Code Writing & Editing

"Fix the pagination bug in the user listing API"
→ OpenHands reads the code, identifies the off-by-one error, writes the fix, runs tests

Command Execution

"Set up a PostgreSQL database with the schema from schema.sql"
→ OpenHands runs Docker commands, creates the database, applies migrations

Web Browsing

"Research the Stripe API docs and implement webhook handling"
→ OpenHands browses stripe.com/docs, reads the webhook guide, implements the handler

Full-Stack Development

"Build a REST API for a todo app with CRUD endpoints, validation, and tests"
→ OpenHands scaffolds the project, writes endpoints, adds validation, creates tests

Architecture

Sandboxed Execution

All code runs inside a Docker container:

  • Network access controlled
  • File system isolated
  • Can install packages safely
  • Full terminal access

Agent Loop

1. Observe: Read the current state (files, terminal, browser)
2. Think: Plan the next action
3. Act: Execute code, run commands, or browse
4. Verify: Check if the task is complete
5. Repeat until done

SWE-bench Results

Agent SWE-bench Verified
OpenHands (Claude) 53%
SWE-agent (Claude) 23%
Aider (Claude) 26%
Devin 14% (original benchmark)

Web UI

Full-featured interface with:

  • Chat panel for task assignment
  • Code editor with live preview
  • Terminal output
  • Browser view (for web tasks)
  • File explorer

Key Stats

  • 45,000+ GitHub stars
  • 53% SWE-bench Verified (highest open-source)
  • Sandboxed Docker execution
  • Web browsing capability
  • Full-featured web UI

FAQ

Q: What is OpenHands? A: OpenHands is an open-source autonomous AI developer that writes code, runs commands, browses the web, and resolves real GitHub issues — achieving 53% on SWE-bench Verified.

Q: Is OpenHands free? A: Yes, fully open-source under MIT license. You need LLM API keys.

Q: Is it safe to let OpenHands run code? A: Yes, all execution happens inside a sandboxed Docker container with controlled network access.

🙏

Source & Thanks

Created by All Hands AI. Licensed under MIT.

OpenHands — ⭐ 45,000+

Thanks for building the most capable open-source AI developer.

Discussion

Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.

Related Assets

AutoGen — Microsoft Multi-Agent Conversation Framework

Framework by Microsoft Research for building multi-agent conversational AI systems. Agents chat with each other to solve tasks collaboratively. Supports human-in-the-loop and code execution. 40,000+ stars.

Agent Toolkit
◈Home🔍Search👤Me