Multitui

Product Introduction

1. Definition: Multitui is a macOS-native application factory that generates individual, sandboxed terminal apps for Text-Based User Interface (TUI) programs. It transforms command-line tools into first-class macOS applications with native integration.
2. Core Value Proposition: Multitui solves macOS workflow fragmentation by enabling secure, isolated execution of TUIs (e.g., AI coding agents, system monitors) while preserving native macOS features like Dock integration, Spotlight search, and Finder context menus.

Main Features

1. Sandboxed Filesystem Protection:
   • How it works: Leverages macOS’s built-in sandbox-exec to enforce granular filesystem rules. Default configuration denies all write/delete operations outside user-specified paths (e.g., project folders).
   • Technology: Implements real-time deny logging with one-click rule adjustments. Supports custom profiles like --dangerously-skip-permissions for advanced users.
2. Native macOS App Conversion:
   • How it works: Wraps TUI binaries (e.g., lazygit, claude) into .app bundles with dedicated icons, Dock/Spotlight visibility, and document-based workflows.
   • Technology: Integrates with macOS APIs for window management (Rectangle/Magnet compatibility), Finder toolbar plugins, and URL schemes (e.g., claudecode://open?arg=/path).
3. Contextual Workspace Isolation:
   • How it works: Creates per-app environments with isolated shell histories, snippets, and themes. Apps retain individual recent documents and support AppleScript/CLI invocation.
   • Technology: Uses macOS’s native document model for session persistence and offers per-app style customization (e.g., red backgrounds for production environments).

Problems Solved

1. Pain Point: Prevents untrusted TUIs (e.g., AI code generators like Claude/Codex) from accidentally modifying/deleting system files. Mitigates risks of malicious or buggy scripts.
2. Target Audience:
   • macOS power users (developers, sysadmins) running CLI tools daily.
   • AI/ML engineers using code-generation agents.
   • Security-conscious professionals handling sensitive data.
3. Use Cases:
   • Securely testing AI-generated code in isolated environments.
   • Running lazydocker or htop as always-on-top system monitors.
   • Editing SQLite databases via harlequin with double-click file support.

Unique Advantages

1. Differentiation: Unlike containerized solutions (Docker/VMs), Multitui requires no environment replication, using existing CLI tools natively. Outperforms basic terminal multiplexers by adding macOS-native UX and security.
2. Key Innovation: Automated sandbox rule management via live deny logs and one-click allowlists—replacing manual sandbox-exec profile creation. Project-specific app instances minimize attack surfaces.

Frequently Asked Questions (FAQ)

1. Does Multitui work with Little Snitch?
   Yes, Multitui integrates with Little Snitch for per-app network control, complementing its filesystem sandboxing.
2. Can I convert any CLI tool to a macOS app?
   Multitui supports any TUI (e.g., nano, jqp, gemini) via executable wrapping, with configurable arguments and environment variables.
3. How does sandboxing protect against data loss?
   Default rules block all writes/deletes outside /tmp and user-allowed paths. Denied actions are logged in real-time for auditing.
4. Is Multitui compatible with macOS Sonoma?
   Yes, it requires macOS 15+ (Sonoma or later) and supports Apple Silicon/Intel architectures.
5. Does Multitui track user activity?
   No, the app contacts its server only for update checks, with zero telemetry or data collection.