OpenChamber: An Open-Source AI Coding Workspace Built on OpenCode

What OpenChamber Is and Who Uses It

OpenChamber is an open-source workspace for running and reviewing AI coding work. It wraps the OpenCode coding agent with a layer that manages sessions, devices, and review workflows. The README explains the relationship: OpenChamber uses OpenCode to run coding agents because it is open source, has a solid API, and is easy to extend, but OpenChamber handles the rest of the workflow: deciding what to try, keeping the agent on track, reviewing the result, connecting from another device, and shipping the change.

The project is an independent product by Bohdan Triapitsyn and is not affiliated with the OpenCode team. The package manager is bun 1.4.2, and the monorepo includes packages for web, UI, desktop (Electron), VS Code extension, mobile, and SDK targets. The MIT license applies to the repository, and the current version is 2.0.3 as of its most recent release on 2026-09-28.

The development pace is high. The v2.0 line introduced Sessions in work and permission modes (v2.0.3), a cleaner chat timeline (v2.0.2), and 2.0 fixes and quality-of-life upgrades (v2.0.1), all released within four days. The last push to the repository was on 2026-09-27. Users following the project should monitor the CHANGELOG.md for breaking changes between minor versions.

Surfaces: Desktop, VS Code, iOS, Android, and Web

OpenChamber targets five surfaces:

The Desktop app runs on macOS, Windows, and Linux, with multiple windows, Mini Chat, remote machine connections, SSH, and native notifications. It bundles the matching OpenCode CLI, so no separate installation is required.

The VS Code extension keeps sessions beside code, lets users send selections to the agent, open results in the editor, and compare parallel runs. Install it from the Visual Studio Marketplace by searching "OpenChamber" in Extensions.

The Web and PWA surface opens the workspace in a browser or installs it as a Progressive Web App with background notifications. The CLI/Server deployment uses the same codebase.

The iOS and Android apps let users review and steer work away from a desk, receive completion alerts, and use the terminal with touch controls.

All surfaces share the same projects and sessions. Checking progress, answering questions, reviewing changes, or reattaching to a running terminal is available from any surface.

Installing OpenChamber: Desktop, CLI, and Docker

For the Desktop app on macOS, Windows, or Linux: download the latest release from the GitHub Releases page. Linux releases are AppImages for x86_64 and ARM64. After download, make the AppImage executable and run it:

chmod +x OpenChamber-*.AppImage
./OpenChamber-*.AppImage

Linux AppImages require FUSE (libfuse.so.2). Without FUSE, the environment variable `APPIMAGE_EXTRACT_AND_RUN=1` is needed.

For the CLI, Web, and PWA: requires Node.js 22 or later and an installed OpenCode CLI. Install OpenChamber with:

curl -fsSL https://raw.githubusercontent.com/openchamber/openchamber/main/scripts/install.sh | bash
openchamber --ui-password be-creative-here

Common CLI operations after install:

openchamber status
openchamber connect-url --qr
openchamber tunnel start --provider cloudflare --mode quick --qr
openchamber startup enable
openchamber logs
openchamber stop
openchamber update

For Docker: the compose file maps port 3000 and requires the OPENCHAMBER_UI_PASSWORD environment variable. The compose file refuses to start if the password is not set. The Dockerfile installs `@opencode/[email protected]` as part of the build.

Session Goals, Multi-Run, and Fusion

Session Goals set a finish condition for an agent session. OpenChamber checks the result after every turn and keeps the agent working until the goal is met, the agent is blocked, or the configured turn limit is reached. Sessions can continue after the user closes the app.

Multi-run assigns the same task to up to five models simultaneously, each in its own session and optionally its own Git worktree. The result is a side-by-side view of what each model actually built, not a comparison of explanations. Users can select the best result or use Fusion to combine the strongest parts into a new session.

Fusion is described in the README as combining the strongest parts from parallel runs into a new session. The README does not describe the exact mechanism; the Fusion documentation is in the guides at packages/docs/content/docs/multi-run.mdx.

These three features address a specific problem in AI coding work: the agent's first attempt at a complex task is not always the best one, and comparing multiple approaches requires managing multiple sessions manually without OpenChamber.

Changes Walkthrough, Preview, and GitHub Integration

Changes Walkthrough converts a large diff into an AI-guided tour of the change. It groups related edits into steps, orders them in a logical sequence, and explains how the pieces fit together. This is aimed at the code review phase, not the coding phase.

Preview opens the user's app beside the conversation. Pointing at an element sends the agent its screenshot, styles, position, and browser errors. The desktop app can do the same with any web page in its built-in browser. This replaces describing a visual bug in text with sending the agent the exact element state.

GitHub integration lets users start a session from a GitHub issue or pull request with its context attached. Failed checks or review comments can be sent back to the agent, and the pull request can be updated or merged from within OpenChamber. The guides directory documents this workflow at packages/docs/content/docs/github.mdx.

Private Relay and Secure Remote Access

Private Relay connects a remote device to an OpenChamber instance using a one-time QR code without opening ports or exposing a public server. The connection is end-to-end encrypted and can be revoked at any time.

Alternative access methods are documented alongside Private Relay: direct connections, LAN and VPN access, Cloudflare and Ngrok tunnels, and SSH. The docker-compose.yml shows Cloudflare tunnel configuration through environment variables: OPENCHAMBER_TUNNEL_PROVIDER, OPENCHAMBER_TUNNEL_MODE, OPENCHAMBER_TUNNEL_HOSTNAME, and OPENCHAMBER_TUNNEL_TOKEN.

The README notes that OpenChamber binds to localhost by default. The `--lan` flag enables LAN binding and the `--ui-password` flag protects browser access. In Docker, the compose file binds to all interfaces for port mapping, making the UI password requirement mandatory rather than optional.

A reverse proxy guide is available at docs/REVERSE_PROXY.md for teams putting OpenChamber behind a reverse proxy.

OpenChamber vs Using the OpenCode CLI Directly

The OpenCode CLI runs coding agents from the command line without a workspace layer. It is the foundation that OpenChamber builds on. Using the OpenCode CLI directly suits developers who work in a single terminal session on a single machine and do not need cross-device access, session persistence across app restarts, or visual diff review tools.

OpenChamber adds cost and complexity: it requires a server process running alongside the OpenCode CLI, a bun-based monorepo for building from source, and the OPENCHAMBER_UI_PASSWORD environment variable before any network-accessible deployment. For a single developer working locally, the OpenCode CLI is sufficient.

For developers who frequently switch between a desktop workstation, a laptop, and a mobile device, or who want to run multiple agent sessions in parallel and compare their results, OpenChamber adds capabilities the CLI does not have. The multi-run feature in particular has no equivalent in the base CLI.

The README states that OpenChamber is not affiliated with the OpenCode team and was built independently. This means the project tracks OpenCode's API but makes its own decisions about release timing and feature priorities.

Editorial conclusion

Developers who run OpenCode agents and want persistent sessions across Desktop, VS Code, iOS, and Android, with multi-run parallel comparisons and Private Relay remote access, should try OpenChamber. The project is MIT-licensed and is an independent product not affiliated with the OpenCode team. The minimum runtime is Node.js 22, and the Docker deployment requires the OPENCHAMBER_UI_PASSWORD environment variable before the container will start. The latest release is v2.0.3 as of 2026-09-28, and the project pushed on 2026-09-27, indicating rapid iteration. Developers who only need the OpenCode agent itself without a workspace layer can use the OpenCode CLI directly without OpenChamber.

Frequently asked questions

Official sources

1. License: MIT
2. openchamber/openchamber on GitHub
3. Project website
4. README
5. Releases