Summary

Warden is a native macOS AI chat client designed for speed, privacy, and seamless integration with the Apple ecosystem. It targets macOS users seeking a performant, battery-efficient alternative to web-wrapper AI applications, offering a premium user experience without compromising data privacy.

How It Works

Built entirely with SwiftUI and 100% native code, Warden eschews Electron or web views for a lightweight, responsive application. This native approach ensures instant launch times, minimal RAM usage (under 150MB), and optimized battery efficiency on Apple Silicon. Data processing is designed to remain on-device, enhancing user privacy by only sending requests to user-chosen AI providers.

Quick Start & Requirements

• Installation: Recommended via Homebrew (brew install --cask SidhuK/warden/warden). Manual download from the releases page is also available. Developers can build from source by cloning the repository and running the Xcode project.
• Prerequisites: macOS. Xcode is required for building from source.
• Resource Footprint: Minimal RAM usage (<150MB).
• Documentation: Download for macOS, View Website.

Highlighted Details

• Broad AI Integration: Supports OpenAI, Anthropic, Gemini, Perplexity, OpenRouter via API keys; local models via Ollama and LM Studio.
• Productivity Features: Organize chats into Projects & Workspaces; create custom AI Assistants (Personas); enable File & PDF Chat.
• Advanced Capabilities: Beta Multi-Model Compare for response evaluation; real-time web search with citations; native code execution and syntax highlighting for developers.
• Native macOS Experience: Menu bar access; fluid animations; liquid glass support for MacOS 26.

Maintenance & Community

Developed by Karat Sidhu. Support is encouraged via starring the repository or financial contributions. A promotional offer for Wispr Flow is included. Community interaction channels like Discord or Slack are not explicitly detailed in the README.

Licensing & Compatibility

Licensed under the Apache 2.0 license, permitting commercial use and integration into closed-source projects. The application is exclusively compatible with macOS.

Limitations & Caveats

The "Multi-Model Compare" feature is currently in beta. The application is limited to the macOS platform. Users on versions prior to 0.9 must manually download the latest release to enable automatic updates.