andrew-waters/orchard

A native UI for Apple Containers and (o)MLX sandboxes, written in swift as a replacement for docker desktop

What it solves

Orchard provides a native macOS desktop interface for managing Apple's container tooling. It simplifies the complex process of running containers, persistent Linux VMs (container machines), and local Kubernetes clusters, while specifically adding a streamlined workflow for running local AI models and sandboxed agents.

How it works

Built with Swift and SwiftUI, Orchard communicates with the container daemon via Apple's native XPC API rather than shelling out to a CLI. This allows for real-time resource monitoring, structured data handling, and deep integration with macOS. For AI workloads, it bridges local MLX model servers (like Ollama or LM Studio) to containers, automatically injecting the necessary network endpoints so that containerized apps can use the OpenAI SDK to access the local GPU.

Who it’s for

Developers and AI engineers on macOS who want a high-performance, isolated environment for running containers and local LLMs without relying on heavyweight Electron-based tools or proprietary runtimes.

Highlights

  • Local AI Sandboxes: Run AI agents behind a hypervisor boundary with a local model endpoint and a kill-switch to prevent API key leaks.
  • Native Performance: Uses Apple's Virtualization.framework for sub-second startup times and kernel isolation.
  • Container Machines: Full management of persistent, stateful Linux VMs with live CPU, memory, and disk usage stats.
  • Kubernetes Integration: One-click creation and management of local k8s clusters via the container k8s plugin.
  • Command Palette: A fuzzy-search interface (Cmd+K) to quickly jump to resources or execute actions like "stop db".
  • Multi-pane Log Viewer: Stream and filter logs from multiple containers side-by-side with color coding.

Related

  • Project
    arcboxlabs/arcbox

    An open-source container and VM runtime for macOS written in Rust that provides a drop-in Docker replacement and isolated microVM sandboxes for AI agents.

  • Project
    missuo/herdrm

    A native macOS console for herdr that allows users to manage and interact with AI coding agents across local and remote machines via a full TUI interface.

  • Project
    osaurus-ai/osaurus

    A native macOS AI harness that provides local agents, memory, and tool execution in a secure sandbox, allowing users to maintain their AI identity and context privately on-device.

  • Project
    webcoyote/sandvault

    A macOS tool that runs AI agents and shell commands in a sandboxed user account to provide isolation without the overhead of virtual machines.

  • Project
    Octane0411/open-vibe-island

    Open Island is a free, native macOS app that lives in the notch or menu bar and provides a live dashboard for AI coding agents (Claude Code, Codex, Cursor, Gemini‑CLI, etc.). It shows session status, handles permission prompts, displays usage stats, and lets you jump back to the exact terminal or IDE pane with a click. Integration is hook‑based: agents emit JSON events, a tiny CLI forwards them via Unix socket to the app, which updates its UI. Supports 13 agents and 15+ terminals/IDEs, runs locally with no telemetry, and is distributed as a signed DMG or Homebrew cask.