Mudroom

Mudroom

Run coding agents in a VM, review changes like a PR

A free, open-source sandbox that runs coding agents in a Linux micro-VM on a project copy, so you review diffs before anything reaches your files.

Mudroom homepage

Mudroom homepage

1 / 2

Overview

What is Mudroom?

Run coding agents in a VM, review changes like a PR

Mudroom is a free, open-source sandbox for developers who run coding agents with permission prompts off and want to see what the agent did before it touches their real project.

Each session clones the project, including untracked and ignored files, and boots a Linux micro-VM with only that clone inside. On a Mac the micro-VM uses Apple's container runtime; on Linux and Windows it uses Docker or Podman. The sandbox has no DNS and no route out by default, and the only exit is a proxy on the user's machine that allows the agent's API hosts plus the hosts added, and logs every connection.

When the agent finishes, Mudroom shows the diff like a pull request. Apply everything, some files or single hunks, and undo any apply later. If a file changed in the project since the session started, Mudroom flags the conflict and will not overwrite it. Mudroom is MIT licensed, with no account and no telemetry.

Platforms and languages

Mudroom Availability

Platforms

MacosWindowsLinux

Languages

English

Capabilities

Mudroom Key Features

Linux micro-VM per session

Runs the agent in a Linux micro-VM on a copy-on-write clone of the project, so the real folder is never mounted.

Network allowlist

Blocks DNS and outbound routes by default, with a proxy allowing only the agent's API hosts and added hosts, logging every connection.

PR-style review

Review the diff like a pull request and apply everything, some files or single hunks, with undo for any apply.

Works with many agents

Supports Claude Code, Codex, Gemini CLI, opencode and Aider, or any command, with local models through Ollama or LM Studio.

App and CLI

Offers a native Swift app for macOS and a command-line tool for Linux, Windows and Intel Macs.

Best for

Who uses Mudroom?

Developers running coding agents such as Claude Code and Codex

Run agents safely in a VM and review their changes before applying them

Plans and access

Mudroom Pricing

Free

Free and open source (MIT); no account or subscription

No free trial listed

Common questions

Mudroom FAQs

Is Mudroom free?

Yes. Mudroom is free and MIT licensed, with no account, no telemetry and no subscription.

Does Mudroom change my real project files?

No. The agent works on a copy in a micro-VM, and nothing reaches your real folder until you apply the diff.

Which agents does Mudroom support?

Mudroom supports Claude Code, Codex, Gemini CLI, opencode and Aider, or any command, with local models through Ollama or LM Studio.

Which platforms does Mudroom run on?

Mudroom offers a native macOS app on Apple silicon, plus a CLI for Linux, Windows and Intel Macs.

Reviews

See what the community thinks and share your experience.

—

Based on 0 ratings

Rating distribution

0
0
0
0
0

Leave a review

Sign in to rate

Community reviews

No reviews

No written reviews yet.

Explore by category

AI Integration Automation

View all AI Integration Automation websites

Explore by category

AI Testing Tools

View all AI Testing Tools websites
StateWeave preview
StateWeave

Stateful sandboxes that clone your apps so AI agents can be tested on a replica of your full system before production.

AI Testing Tools
Seldon preview
Seldon

QA agents that test your product like real customers: describe tests in plain English and get replays, screenshots and reproduction steps for failures.

AI Testing Tools
Glasshand preview
Glasshand

Rents real iPhones and Macs in the cloud so AI agents can use apps, test software, and handle everyday phone tasks like a person.

AI Testing Tools