Devtoolwatch BoardGet it every Monday

OpenHands review

OpenHands Agent Canvas is a self-hosted control center for running coding agents and automations across local, remote, and cloud backends.

Updated Nov 1, 2026. Review written Oct 22, 2026 from the project's own sources; usage data refreshed every Monday.

TL;DR

OpenHands Agent Canvas is a self-hosted developer control center for coding agents and automations. It suits developers and teams who want agents running locally, in Docker, on VMs, or on servers. It runs the open source OpenHands agent by default and can also run Claude Code, Codex, Gemini CLI, Pi, OpenCode, or any ACP-compatible agent.

OpenHands at a glance

Category
Coding agents
Maintained by
OpenHands (organization, @OpenHands on GitHub)
License
MIT
Language
TypeScript, JavaScript, Python
Repository
github.com/OpenHands/OpenHands
Website
openhands.dev
PyPI package
openhands-ai 1.11.0
Latest release
v1.29.0 on GitHub, Oct 20, 2026
First published
Mar 13, 2024

What OpenHands does

  • OpenHands runs coding agents locally, in Docker containers, on VMs, or within company infrastructure, and you can switch between these backends from one frontend.
  • OpenHands runs third-party agents such as Claude Code, Codex, Gemini CLI, Pi, and OpenCode, plus any agent using Agent-Client Protocol (ACP).
  • OpenHands creates automations that run on a schedule or in response to webhook events.
  • OpenHands automations integrate with services such as Slack, GitHub, Linear, and Notion, for example to decompose GitHub issues into tasks.
  • OpenHands works with any LLM through configurable model settings.
  • OpenHands can run each conversation in its own Docker container, with its own Agent Server and tools, so several agents can run concurrently.
  • OpenHands lets the agent request a secret through an in-conversation modal.

How it works

OpenHands Agent Canvas is a frontend powered by the OpenHands Agent Server, a REST API for running multiple agents on one machine. Each Agent Server runs on one host and port. Canvas can connect to several servers and switch between them. An optional Automation Server runs agents on schedules or in response to events. You can also run agents on OpenHands Cloud or OpenHands Enterprise infrastructure.

How to install and start using OpenHands

With npm

npm install -g @openhands/agent-canvas

With multiple Docker sandboxes

OH_CONVERSATION_RUNTIME=docker agent-canvas

From source

git clone https://github.com/OpenHands/OpenHands.git

Requirements

  • Node.js 24 or later and uv for the npm and source installs.
  • Docker Desktop on macOS or Windows, or Docker Engine or Docker Desktop on Linux, for the Docker sandbox options.
  • A host directory set as PROJECTS_PATH, holding the projects the agent may access, for the single Docker sandbox.
  • Python 3.12 or later but below 3.14 for the openhands-ai PyPI package.

First steps

  1. Install OpenHands Agent Canvas globally with npm, then run the agent-canvas command to start the full local stack.
  2. Open http://localhost:8000 for the npm or source launchers, or http://localhost:8000/canvas for the Docker image.
  3. Start a new conversation in the UI, and add more backends directly from the UI.

Who OpenHands is for

  • Developers who want to self-host coding agents on a laptop, a dedicated machine, or a cloud server.
  • Teams that want to share an Agent Server for code review and dependency updates.
  • Engineers automating workflows triggered from Slack, GitHub, Linear, Jira, or schedules.
  • Developers who want to run several agents, such as Claude Code or Codex, from one control center.

Limitations to know

  • OpenHands installed without a sandbox runs the agent server directly on your machine, which gives the agent full access to your filesystem.
  • OpenHands conversations that share a host workspace also share its files. Use separate directories or worktrees to avoid conflicting edits.
  • OpenHands multi-sandbox mode isolates conversation execution only. It does not move Canvas or the automation service into a sandbox.
  • OpenHands internet-facing installs need security hardening per the self-hosting guide, including API key entry instead of an injected session key.

Is OpenHands growing? Usage data

In the week of Oct 25–31, 2026, OpenHands ranked #284 on the Devtoolwatch Board and #90 in Coding agents, with a momentum score of 33 out of 100. Its main package had 82.5k downloads that week, −53% against its own 4-week average.

New GitHub stars per week for OpenHands12 weeks from Aug 9, 2026: 697, 703, 810, 740, 1.5k, 913, 722, 745, 661, 718, 690, 670 new stars, an average of 794 a week.Aug 9week of Oct 25
New GitHub stars per week for the last 12 weeks, from GitHub's stargazer history (today's stargazers by the week they starred).
Show the weekly numbers
New GitHub stars per week for OpenHands
Week ofNew stars
Aug 9, 2026697
Aug 16, 2026703
Aug 23, 2026810
Aug 30, 2026740
Sep 6, 20261.5k
Sep 13, 2026913
Sep 20, 2026722
Sep 27, 2026745
Oct 4, 2026661
Oct 11, 2026718
Oct 18, 2026690
Oct 25, 2026670

The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.

OpenHands alternatives

Other AI coding agents on the Board in the week of Oct 25–31, 2026, by momentum:

Coding agents ranked by momentum, the week of Oct 25–31, 2026
RankToolStarsDownloads/wk
#1CindyCindy is an open-source AI agent app for desktop and mobile that combines Claude Code, Codex and Pi across your connected devices.8.8k–
#2OpenMausBotOpenMausBot is an open-source desktop chat app where each contact is a local Claude, Codex, or Grok CLI agent with its own computer.9.6k–
#3OrcaOrca is a desktop agent development environment for running Claude Code, Codex, and other CLI coding agents in parallel, each in its own…111.7k–
#4OpenBotOpenBot is an open-source, self-hosted template for AI coworkers that each get their own browser, files and granted tools, with every…9.1k–
#5Agent ReachAgent Reach is a Python CLI that lets AI agents search and read platforms such as Twitter, Reddit, YouTube, GitHub, Bilibili and…107.2k–
#6Skill RecorderSkill Recorder is a desktop app that records your on-screen work session and turns it into a reusable agent Skill or Automation.5.6k–

All AI coding agents, ranked

Recent OpenHands releases

OpenHands FAQ

Is OpenHands free and open source?

Yes. OpenHands's source code is on GitHub (OpenHands/OpenHands) under the MIT license, so you can use it for free. Its website also links to a pricing page (https://openhands.dev/pricing) for any paid plans.

How do I install OpenHands?

Run npm install -g @openhands/agent-canvas. The openhands-ai package needs Python 3.12 or newer, below 3.14. The install section above has the other options from the project's README.

Is OpenHands still maintained?

Yes. The OpenHands/OpenHands repository was last updated on Oct 22, 2026, and the latest release, v1.29.0 on GitHub, came out on Oct 20, 2026.

How popular is OpenHands?

OpenHands has 92.4k GitHub stars as of Nov 1, 2026. It gained 670 new stars in the week of Oct 25–31, 2026, against a 12-week average of 794 a week. Its main package had 82.5k downloads that week, −53% against its own 4-week average. That made it #90 among AI coding agents on the Devtoolwatch Board.

OpenHands vs OpenCode: how do they compare?

OpenHands and OpenCode are both tracked by Devtoolwatch as AI coding agents. OpenHands has 92.4k GitHub stars and OpenCode has 223.6k. OpenCode is an open source AI coding agent that helps you write and run code in a terminal, desktop app, or IDE extension. OpenCode review.

OpenHands vs Claude Code: how do they compare?

OpenHands and Claude Code are both tracked by Devtoolwatch as AI coding agents. OpenHands has 92.4k GitHub stars and Claude Code has 154.1k. In the week of Oct 25–31, 2026, OpenHands's main package had 82.5k downloads and Claude Code's had 11.6M. Claude Code is Anthropic's agentic coding tool that reads your codebase, edits files, runs commands, and handles git workflows through natural language. Claude Code review.

What are the best alternatives to OpenHands?

Other AI coding agents on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: Cindy (#1), OpenMausBot (#2), Orca (#3), OpenBot (#4), Agent Reach (#5). The alternatives table above has their numbers.

How does OpenHands work?

OpenHands Agent Canvas is a frontend that connects to one or more OpenHands Agent Servers. Each server is a REST API that runs agents on a host. An optional Automation Server decides when work runs and dispatches conversations to the Agent Server.

Is OpenHands safe to run on my machine?

OpenHands without a sandbox gives the agent full access to your filesystem. The Docker options limit access to the projects under PROJECTS_PATH or to per-conversation containers. Local listeners bind to 127.0.0.1 by default.

Can OpenHands use other agents like Claude Code or Codex?

Yes. OpenHands Agent Canvas runs the OpenHands agent by default. It can also run Claude Code, Codex, Gemini CLI, Pi, OpenCode, or any ACP-compatible agent.

Which models does OpenHands support?

OpenHands can be used with any LLM, which you configure through its LLM settings.

Can OpenHands keep running when my laptop is off?

Yes, if you run OpenHands on a cloud server. A server setup keeps agents running when your laptop is shut and makes it easier to trigger them from services like Slack, GitHub, and Datadog.

Sources

  1. OpenHands/OpenHands on GitHub
  2. OpenHands README
  3. openhands-ai on PyPI
  4. OpenHands website
  5. Release v1.29.0 (Oct 20, 2026)
  6. Release v1.28.1 (Oct 19, 2026)
  7. Release v1.28.0 (Oct 19, 2026)
  8. Devtoolwatch Board, the week of Oct 25–31, 2026

OpenHands README badge

Maintainers can show the tool's weekly Board rank in their README:

Devtoolwatch badge for OpenHands

[![Devtoolwatch](https://index.devtoolwatch.com/badge/openhands.svg)](https://index.devtoolwatch.com/tools/openhands/)

How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].