A Native MacOS Chatbot that works for every AI LLM

Warden is a minimalist, simple and beautiful macOS AI chat app, that supports most AI providers: ChatGPT, Anthropic (Claude), xAI (Grok), Google Gemini, Perplexity, Groq, Local LLMs through Ollama, OpenRouter, and almost any OpenAI-compatible APIs.

Look, I needed an AI app. Not just any AI app, but one I'd actually want to fire up daily. And let's be super clear: absolutely no Electron bloat, no web wrappers pretending to be apps. I wanted a proper, native Mac experience, something beautiful, with personality, that truly belonged on my desktop. I wanted the app to feel like something apple would create, it didn't need to be the fastest app out there (and its not, although its fast enough for most people), but there are little things in this app that make me smile when I use it.

Subtle animations and effects throughout the app:

Like swipe gestures:

and a bunch of different subtle design choices that make this lightweight app a joy to work with.

Lets talk about some actual features now:

• Completely Native app, built with SwiftUI
• Zero telemetry, no data collected from you, all your data remains on your Mac and can be easily exported into an easy-to-read JSON file
• Lightweight app, compared to Electron based alternatives
• Supports these LLM APIs:
  • Openrouter
  • OpenAI
  • Anthropic (Claude)
  • xAI (Grok)
  • Groq
  • Perplexity
  • Ollama
  • Deepseek
  • It should also support most other OpenAI compatible APIs
• Fast chat responses, across all APIs

Even More Features:

• Ability to create custom chat assistants, set their system prompt and temperature

  

  
  
• Ability to control context windows on a per message basis

  

• Beautiful syntax-highlighted code blocks for most coding languages
• Ability to attach images to prompts
• Artifacts: Ability to run simple HTML/CSS code directly in the app

  

  
  
• Easily switch between light and dark mode
• Customize every chat with a different instruction, system prompt
• Native interface/UI that actually follows the MacOS design language
• Copy chats and code right from the conveniently located copy buttons
• Lightweight, this version of the app is less than 20 MB, and in my testing doesn't tend to use more than 150 MB of RAM for most chats (YMMV,ofc)
  

Installation:

• This app is not signed/notarized by apple developer accounts
• Make sure you go to your mac's privacy and security settings and turn on install from unknown developers
• If you have problems with installation, contact me on twitter or github

Updates:

Right now the app doesn't connect to my end, so even though I have integrated the Sparkle Swift framework, there will be no updates pushed automatically, therefore, please check the github repo where I will be publishing updates as I work on them:

https://github.com/SidhuK/WardenApp

More Screenshots

Future Development

This is a project that I plan to continue on working on in my spare time, development will vary depending on how much free time I have. There are some features I plan to add to this app, but they will be few and far, because I don't want to add features that I personally consider gimmicks for most AI apps.

Some future major features I plan to add include:

• MCP Support
• Apple Shortcuts support
• Web searching capabilities (can be done now, but only using Perplexity)
• Image Generation Abilities
• Chatting with multiple AI models at the same time

Credits:

This project is forked and heavily inspired from the MacAI app, created by the talented Renat, huge props to them for making the source code of their app open source and Apache-licensed.

Disclaimer:

I bear no responsibility for any of your data loss, etc. This is my first ever swift app, so please bear with me as I continue to work on getting rid of bugs etc.

Donations

Again, this app is completely free to use, however, if you would like to contribute to its further development feel free to either buy it from here, or click on my coffee link below