Collab Code

Collaborative coding rooms for classrooms, tutoring sessions, and pair programming.

Share a room link, edit the same workspace in Monaco, use one shared terminal session, manage files together, and run Java or Python from the browser through a lightweight relay server.

Why Collab Code

Collab Code is built for the "open a link and start coding" workflow. It is meant to feel more like a shared coding room than a demo editor: multiple files, a real explorer, a shared terminal, search, diagnostics, and live collaboration all sit in the same browser-first workspace.

It is especially suited to tutoring, CS education, pair programming, and collaborative walkthroughs where setup friction gets in the way.

Highlights

Collaboration	Workspace	Execution
Live cursors, selections, peer presence, room sharing, shared terminal session	Multi-file explorer, tabs, search/replace, drag-and-drop import, multi-select file actions	Interactive Java and Python run flow, inline diagnostics, shared terminal I/O

• Collaborative editor - Monaco + Yjs with live remote cursors, selections, and per-file awareness.
• Collaborative terminal - peers share the same terminal transcript, prompt state, working directory, and run session.
• Real workspace - folders, tabs, file icons, context menus, workspace search, export, and drag-and-drop file management.
• Integrated terminal tools - browse and modify the virtual filesystem without leaving the app.
• Classroom-friendly sharing - create or join a room instantly with no account flow.
• Offline-friendly persistence - IndexedDB keeps local workspace state around between reconnects.
• Responsive UI - explorer, search, editor, and terminal all adapt across desktop and smaller screens.

Current Scope

• Runnable languages today: Java and Python
• Editor/file support: Java, TypeScript, JavaScript, Python, JSON, HTML, CSS, Markdown, C, C++, XML, and SQL
• Best fit: classrooms, tutoring sessions, collaborative exercises, code reviews, and quick pair-programming rooms

Quick Start

Try it live

Open the deployed app:

https://ianskelskey.github.io/collab-code

Run it locally

npm install
npm run dev:all

Then open:

http://localhost:5173/collab-code/

For local development details, environment variables, architecture notes, deployment, and contributor workflow, see CONTRIBUTING.md.

Execution Safety

Browser-triggered Java and Python execution runs arbitrary user code on the relay host, so deployment defaults matter.

• The Docker image is configured to require sandboxed execution and drops Java/Python child processes to a dedicated low-privilege OS user.
• Python execution now runs inside a fresh per-run virtual environment. If the workspace includes a requirements.txt next to the entry file or in one of its parent folders, those packages are installed into that isolated environment instead of the server's global Python installation.
• If the server is started as root without EXEC_SANDBOX_UID and EXEC_SANDBOX_GID, execution is now disabled by default. You can explicitly override that with EXEC_ALLOW_UNSANDBOXED_ROOT=1, but that is not recommended.
• Local development on platforms without POSIX uid/gid privilege dropping can still run unsandboxed unless you require sandboxing with EXEC_REQUIRE_SANDBOX=1.
• The app's Help -> About tab and the server health endpoint report whether execution is Sandboxed, Unsandboxed, or Disabled.

This is a hardening step, not a full container-per-run sandbox. If you expose execution to untrusted users, isolate the relay further at the deployment level as well.

What You Can Do

• Create or join a room from the landing page in a few seconds.
• Edit the same workspace with other people in real time.
• Share one terminal session with other peers, including cd, command history, terminal output, and interactive Java/Python stdin/stdout/stderr.
• Search and replace across the entire workspace.
• Manage files with the explorer, drag-and-drop, multi-select actions, and the built-in terminal.
• Compile and run Java projects, or execute Python scripts inside an isolated virtual environment with streamed stdin, stdout, and stderr in the shared terminal.
• Export a file or the whole workspace when you are done.

Built With

React, Vite, Monaco Editor, Yjs, xterm.js, and a small Node/WebSocket relay server.

Contributing

Bug reports, feature ideas, and pull requests are welcome. Start with CONTRIBUTING.md.

License

Released under the MIT License. You're free to use, modify, and redistribute Collab Code, including commercially, as long as the license and copyright notice are preserved.