Artwork by Halyna
9.6K stars on GitHub

An agentic development environment.

It’s open source and works with the models you already use. Start an agent on one machine, then pick up the same session from another.

OpenChamber desktop interface

What runs where.

OpenChamber runs as an instance on a machine you control. The desktop app runs one on your computer and connects to others on remote machines. Your phone and a browser reach any of them over the internet.

Instance

Mac, Linux box, or server

  • Starts on your Mac, your Linux box, or a server you own
  • Runs the agents and keeps every session, even after you close the app
  • Listens on localhost by default. One flag opens it to your network
Local and remote

Desktop app

  • Runs its own instance and connects to remote ones as well. Switch between them from the app
  • macOS, Windows, and Linux
  • Terminal, file browser, notes, todos, and reusable actions next to the chat
  • Provider limits, token usage, and cost per session
  • Also available as an extension for VS Code, Cursor, Windsurf, and Trae
OpenChamber desktop app
Remote

Phone and browser

  • A native mobile app, now in beta, and a PWA
  • Private Relay carries an end-to-end encrypted connection without an open port
  • Pair a phone with a QR code once. Revoke it from the app whenever you like
  • A tunnel with a UI password also works if plain browser access is enough
OpenChamber session list on mobile
OpenChamber conversation on mobile

Start, run, review.

Every task goes through the same three stages. This is what OpenChamber adds at each one.

  1. Start

    Start from an issue, a goal, or a schedule.

    • From GitHub or Linear

      Start from a GitHub issue or pull request, or a Linear ticket. The agent gets the thread and the diff with it. GitLab is in the works.

    • Session Goals

      Set the outcome. After every turn OpenChamber checks progress and continues until the goal is done, blocked, or over the limit you set.

    • Scheduled work

      Run a task once at a set time or on a recurring schedule. With a goal attached, each run keeps going until the work is done.

  2. Run

    Run the same task on several models.

    • Multi-run

      One task across up to five models, in parallel, each in its own worktree.

    • Fusion

      Keep the best result, or fuse the strongest parts of several into one.

    • Browser Use

      Your running app in the built-in browser. You click through it, the agent does too, sees the errors, and checks its own work.

  3. Review

    Review the change as ordered steps.

    • Changes Walkthrough

      A large diff, grouped into ordered steps that explain how the change fits together.

    • Failed checks go back

      Send a red CI run to the agent that wrote the code. It comes back with the fix.

    • Merge without leaving

      Commit, push, open the PR, answer review comments, and merge from the same window the work happened in.

Privacy & Data

OpenChamber collects nothing about your work and sends nothing anywhere. Code and session content stay on your machine. When the workspace leaves localhost, you gate remote access with a UI password or a paired device.

Code and sessions stay with you

OpenChamber does not collect project names, paths, prompts, code, diffs, or session content.

Remote access is protected

A UI password gates browser access. You can rotate or revoke tunnel links the moment they should stop working.

Private Relay, no open ports

Pair a device with a one-time QR code and connect through Private Relay without opening a port or exposing a public server. The connection is end-to-end encrypted, and you can revoke it at any time.

Open source, so you can check

OpenChamber is open source. The privacy model is in the code, and anyone can check it.

From people who use it.

Posts we did not ask for.

Image @realchendahuang
I've had a deep experience with openchamber. I think my choice was right. This thing is really useful, smooth and silky. Go try it out quick, I strongly recommend it, no more messing around blindly.
Image @NihilSineTheo
Was looking for a way to run agents from my phone, remotely on my server, other than tmux. Found OpenChamber. It is INSANELY good, highly recommended.
Image @1un4rPunk
Seriously, I've spent almost the whole week trying to find a tool that made sense to orchestrate agents and, as a dev (actually, a devops), openchamber is the only one that felt natural and logical.
Image hkay-dev
I just want to say the app itself is phenomenal. The level of polish is just unreal. I saw OpenChamber back when it was a small project, I think maybe a couple months ago, and now it is unrecognizable. Truly amazing work. The latest release with its streaming feature is everything I could have possibly asked for and more. It is gorgeous and smooth.

OpenChamber FAQ

A machine you control with OpenCode installed. OpenChamber runs on the OpenCode SDK. Install it with curl -fsSL https://opencode.ai/install | bash, then start OpenChamber.
On your Mac, your Linux box, or a server you own. The desktop app runs an instance locally and connects to remote ones too. Your phone and a browser reach any of them through Private Relay.
Pair a device through Private Relay. The connection is end-to-end encrypted, no port is opened, and you can revoke the device at any time. A tunnel with a UI password also works if plain browser access is enough.
Yes. OpenChamber is MIT-licensed, and the code is on GitHub.
Nothing. OpenChamber itself is free. You pay only your model provider, on your own keys.

More questions? Check the documentation

Install it on the machine you already have.

One command on a Mac, a Linux box, or a server. Then open it from wherever you are.