OpenChamber

// Overview

OpenChamber is an open-source workspace for running and reviewing AI coding work. Rather than being an agent itself, it wraps the OpenCode CLI agent in a session manager you can drive from five surfaces — a desktop app for macOS, Windows and Linux, a web/PWA build, a VS Code extension, iOS and Android, and a CLI/server mode. Your projects and sessions stay available when you switch devices or step away, so a session started at a workstation can be steered from a phone.

The workflow it is built around is start work, see what changed, review, release. Session Goals give a session a finish line: OpenChamber checks the result after every turn and keeps the agent working until it completes the goal, gets blocked, or hits the limit you set — including after you close the app. Multi-run gives the same task to up to five models, each in its own session and optionally its own worktree, and Fusion combines the strongest parts of those runs into a new session. Changes Walkthrough turns a large diff into a guided tour that groups related edits into ordered steps.

For remote access, pairing a device with a one-time QR code sets up Private Relay, an end-to-end encrypted connection that needs no open ports and can be revoked at any time; direct connections, LAN/VPN, Cloudflare and Ngrok tunnels, and SSH are also supported. The server binds to localhost by default, and the docs are explicit that `--lan` belongs only on a trusted network and browser access should be protected with `--ui-password`.

// What it does

• Session Goals: the agent keeps working toward a stated outcome across turns, and continues after you close the app
• Multi-run and Fusion: run one task against up to five models in parallel worktrees, then merge the best parts into a new session
• Changes Walkthrough: a large diff replayed as an ordered, explained tour of the change
• Preview: open your running app beside the conversation and point at an element to send the agent its screenshot, styles, position and browser errors
• GitHub flow: start a session from an issue or pull request, send failed checks and review comments back to the agent, then update or merge
• Private Relay, an end-to-end encrypted, revocable device pairing that needs no open ports, plus scheduled and recurring prompts

// Getting started

Desktop bundles the matching OpenCode CLI, so no separate install is needed there; the CLI/Web and VS Code paths use your own OpenCode CLI install and require Node.js 22+.

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

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

openchamber connect-url --qr
openchamber tunnel start --provider cloudflare --mode quick --qr

openchamber startup enable
openchamber status
openchamber logs
openchamber update
openchamber stop

code --install-extension fedaykindev.openchamber

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Hand a task a finish line and let it keep working while you are away, then review the whole change as a guided walkthrough
• Race several models on the same task in separate worktrees and keep whichever result is actually better
• Review and steer an agent from a phone or another machine without exposing a public server
• Turn a GitHub issue into a session, push a pull request, and feed failing checks straight back to the agent

// How OpenChamber compares

OpenChamber alongside other open-source autonomous coding agents tools AI/TLDR tracks, ranked by GitHub stars.