This application allows LLM researchers to interact with self-hosted models through a chat interface. It supports the Ollama API and its ecosystem, enabling users to test prompts and switch between models with conversation context. The app also provides nicely formatted output for Markdown elements like tables, lists, and code.
What We Love
- ๐ฏ Clean, minimalist, lightweight interface
- โก Fast and responsive performance
- ๐ Privacy-focused local LLM support
- ๐ค Voice recognition and text-to-speech capabilities
- โจ Seamless Ollama integration
- ๐ฌ Intuitive chat experience
- ๐ Completely free and open-source
- ๐ง Easy setup and configuration
- ๐ฑ Native Apple app (not Electron)
- โจ๏ธ Keyboard shortcuts and Siri integration
- ๐จ Aesthetic and polished UI design
- ๐ Auto-detection of local models
- ๐ File and image attachment support
Room to Improve
- ๐ Frequent app crashes and freezing
- ๐ Connection drops when app goes to background
- โฑ๏ธ Non-configurable timeout settings
- ๐ Auto-scrolling prevents reading while streaming
- ๐ฏ Input field loses focus during response generation
- ๐ Limited to Ollama, incompatible with LLM Studio
- ๐ฑ iOS features significantly limited
- ๐ No RTL language support
- ๐ค Poor contrast in light mode
- ๐พ No cross-device sync capability
- ๐ผ๏ธ Crashes with large file uploads
- ๐๏ธ Poor quality TTS output
- โจ๏ธ Difficult to hit submit button near audio button
