OpenChamber

OpenChamber is a workspace for developers who want to run AI coding agents on their own computer or server and review the results. It uses OpenCode to run agents and keeps projects and sessions accessible across devices.

• Compare model runs. Give the same task to several models in separate sessions, optionally with separate Git worktrees and branches, then choose a result or combine their work with Fusion.
• Review code and running apps. Changes Walkthrough groups a diff into explained steps, while Preview lets you send an agent a selected page element's screenshot, styles and browser errors.
• Goals and scheduled work. Agents can keep working toward a goal across turns until a check finds it complete or blocked, or a limit stops the session; scheduled prompts can use the same goal checks.
• Issues and review feedback. Start with context from a GitHub issue, pull request or Linear ticket, and send failed checks or review comments back to the agent. Integrations include Figma and Sentry.
• Continue across devices. Check progress, answer agent questions, review edits and reconnect to a running terminal from another device.

The MIT-licensed, open-source desktop app runs on macOS, Windows and Linux. You can also self-host the workspace on a server and access it through a browser or VS Code. Private Relay provides end-to-end encrypted remote access without opening a port. OpenChamber states that it collects no data about your work.