claudia-term
Health Warn
- License — License: MIT
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Low visibility — Only 5 GitHub stars
Code Pass
- Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
- Permissions — No dangerous permissions requested
This tool is a native macOS desktop wrapper for the Claude Code CLI. It provides a graphical interface with better visual styling, real-time streaming, and subagent tracking while retaining full access to your existing CLI configuration, MCPs, and sessions.
Security Assessment
Overall risk: Low. The automated code scan of 12 files found no dangerous patterns, hardcoded secrets, or requests for dangerous permissions. Because it acts as a wrapper around the Claude Code CLI, it inherently executes the shell commands generated by the AI agent. The tool distributes pre-compiled macOS `.dmg` files rather than running purely from auditable TypeScript source, which means you are trusting the developer's build artifacts rather than just the open-source code. No suspicious network requests were identified.
Quality Assessment
The project is actively maintained, with its most recent push happening today. It uses the standard, permissive MIT license. The primary concern is its low visibility; having only 5 GitHub stars indicates a very small user base and limited community oversight. The developer openly notes that this is a personal side-project provided as-is, with no guaranteed support or warranty. While the code itself appears clean, the lack of widespread community testing means bugs or edge-case vulnerabilities might be present.
Verdict
Safe to use, provided you are comfortable running a low-visibility developer's pre-compiled macOS application alongside your existing Claude Code setup.
Better looking Claude Code
Claudia-term
Claude Code, but pretty
Claude Code is powerful. The terminal is not where you want to live. Claudia wraps the CLI in a native macOS app so you keep everything—MCPs, CLAUDE.md files, hooks, session persistence—and gain a visual interface that doesn't hurt to look at.
Why rebuild the agent runtime when you can wrap it?
Requires macOS 12+ and Claude Code CLI installed
Screenshots
Parallel subagents running with real-time status tracking
Plan mode with task panel and detailed plan view
Interactive questions with multiple choice options
Note: Claudia is a personal project, provided as-is with no warranty or guaranteed support. Use at your own risk. Bug reports are welcome, but responses and fixes are best-effort.
Why Claudia?
Claude Code CLI is powerful but terminal-focused. Claudia gives you:
- Proportional fonts and styling — Who wants to stare at monospaced text all day
- Visual tool tracking — See what Claude is doing with collapsible, syntax-highlighted tool blocks
- Real-time streaming — Text and results appear as they're generated
- Desktop integration — Native macOS app, multi-window support, project-aware launching
- All of Claude Code's power — MCPs, skills, hooks, CLAUDE.md, session persistence, prompt caching
Why Claudia instead of Cowork?
Cowork is Anthropic's managed desktop experience—great for users who want zero-config simplicity. Claudia is for developers who already use Claude Code and want more control:
- Fully customizable — UI, workflows, features, and memory are yours to modify
- Per-project configuration — different context, tools, and MCPs for each project directory
- Your setup just works — existing CLAUDE.md files, hooks, and MCP configs carry over
- Direct file system access — no sandboxed VM sitting between you and your files
- Custom visualizations — build data renderers directly into your MCPs
Cowork wraps a sandboxed environment. Claudia wraps Claude Code itself—so you keep everything you've already built.
Features
- Native macOS app — Built with Tauri + SolidJS for fast, lightweight performance
- Rich markdown rendering — Full GitHub-flavored markdown with syntax-highlighted code blocks
- Inline image display — Images from Read tool results display inline, click to view full-size
- Visual tool tracking — Collapsible, syntax-highlighted tool blocks with JSON formatting
- Real-time streaming — Text and results appear as they're generated
- CLI launcher — Run
claudiafrom any directory to use project-specific configs - Multi-instance support — Multiple windows, each in different project directories
- Type-ahead input — Keep typing while waiting for responses
- Smart permissions — Auto-approve, plan mode, or per-tool dialogs
- Extended thinking — Expandable thinking blocks when enabled
- Subagent visualization — Track nested Task agents and their progress
- Session resume — Continue previous conversations with full context
Installation
Download (Recommended)
Download the latest release and drag Claudia.app to your Applications folder:
Requires macOS 12+ and Claude Code CLI installed.
Build from Source
If you prefer to build from source, you'll need:
- Node.js 18+
- Rust toolchain
- Claude Code CLI (
claudecommand available)
git clone https://github.com/JasonBates/claudia.git
cd claudia
npm install
npm run tauri build
cp -R src-tauri/target/release/bundle/macos/Claudia.app /Applications/
CLI Launcher
Install the claudia CLI launcher to open the app from any directory:
./install.sh
Or manually:
cp claudia ~/.local/bin/
chmod +x ~/.local/bin/claudia
Then from any project directory:
cd ~/Code/repos/my-project
claudia
This opens Claudia with that directory as the working directory, picking up project-specific .claude/ configs.
Usage
Keyboard Shortcuts
| Shortcut | Action |
|---|---|
| Escape | Interrupt current response |
| ⌘ , | Open settings |
| ⌘ ⇧ [ | Toggle session sidebar |
| ⌥ T | Toggle thinking display |
| ⌥ L | Focus message input |
| ⌥ Q | Quit application |
Commands
| Command | Description |
|---|---|
/clear |
Clear conversation history (restarts Claude process) |
/resume |
Open sidebar to resume a previous session |
/settings |
Open appearance settings |
/thinking |
Toggle extended thinking display |
/sidebar |
Toggle session sidebar |
/exit /quit /q /x |
Close the application |
! <command> |
Execute bash command directly (e.g., ! ls -la, ! git status) |
Resuming Sessions
- Via sidebar: Press
⌘ ⇧ [or type/resumeto open the session sidebar - Via CLI: Launch with
claudia --resume <session-id> - Browse history: The sidebar shows recent sessions with timestamps and previews
Custom Color Schemes
Claudia includes 6 bundled color schemes (Solarized Dark/Light, Dracula, Nord, One Dark, Gruvbox Dark). You can add more by installing iTerm2 color scheme files.
To install new color schemes:
Create the colors directory if it doesn't exist:
mkdir -p ~/.config/iterm2/colorsCopy
.itermcolorsfiles into the directory:cp "My Theme.itermcolors" ~/.config/iterm2/colors/Open Settings (
⌘ ,) — your new schemes appear under "From iTerm2"
Where to find color schemes:
- iTerm2-Color-Schemes — 250+ schemes in the
schemes/directory - Any
.itermcolorsfile from the web will work
Note: iTerm2 does not need to be installed. Claudia reads the .itermcolors plist format directly.
Advanced Runtime Config
You can override model and runtime binaries in ~/.config/claudia/config.json (or per-project at .claudia/config.json):
{
"claude_model": "opus",
"claude_binary_path": "/opt/homebrew/bin/claude",
"node_binary_path": "/opt/homebrew/bin/node",
"legacy_permission_hook_polling": false
}
claude_model: Passed to Claude CLI as--modelclaude_binary_path: Optional explicit path to Claude CLInode_binary_path: Optional explicit Node path used to launch the bridgelegacy_permission_hook_polling: Enables legacy file-based permission flow (off by default)
Architecture
Claudia wraps the Claude Code CLI to leverage its built-in features (MCPs, skills, hooks, session management, prompt caching) while providing a custom native UI.
┌─────────────────────────────────────────────────────────────────┐
│ Claudia.app (Tauri) │
│ Custom UI + Desktop Integration │
├─────────────────────────────────────────────────────────────────┤
│ Frontend (SolidJS) │ Backend (Rust) │
│ └─ Reactive UI components │ └─ Process management, IPC │
├─────────────────────────────────────────────────────────────────┤
│ sdk-bridge-v2.mjs (Node.js) │
│ └─ Thin event translation layer │
├─────────────────────────────────────────────────────────────────┤
│ Claude Code CLI │
│ └─ Full agent runtime: MCPs, skills, hooks, │
│ sessions, compaction, prompt caching │
└─────────────────────────────────────────────────────────────────┘
For detailed architecture, data flow, event types, and implementation details, see docs/architecture.md.
Development
# Development mode
npm run tauri dev
# Run all tests (608 tests: 113 Rust + 495 TypeScript)
npm run test:all
# TypeScript tests only
npm run test:run
# Rust tests only
npm run test:rust
Documentation
| Document | Description |
|---|---|
| docs/architecture.md | Data flow, event types, state management, testing, design decisions |
| docs/streaming.md | Streaming command runner pattern |
| docs/troubleshooting.md | Common issues, debugging techniques, lessons learned |
| CONTRIBUTING.md | How to contribute to Claudia |
| CHANGELOG.md | Version history and release notes |
License
MIT
Reviews (0)
Sign in to leave a review.
Leave a reviewNo results found