Skip to content

Latest commit

 

History

1,176 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Welcome to Compass Compass Logo

Compass is a modern, open-source Large Language Model (LLM) client designed to provide a seamless AI chat experience across multiple platforms. Built with React Native and Expo, it offers a rich set of features while maintaining high performance and user experience. The goal is to promote private, decentralized AI - whilst also maintaining access to cloud-based AI through API's for those who need or prefer it.

Download from latest release

Try it out here (only for demo)

Compass Desktop      Compass Mobile

Features

  • 🌐 Cross-Platform: Buildable for iOS, Android, Linux and web
  • 🔍 Auto-scanning for Ollama instances on network: Just connect to the same network and Compass will find it with the press of a button.
  • 🤖 Multiple LLM Providers: Support for OpenAI, Anthropic, Ollama, and more
  • 👥 Character System: Built-in characters and support for custom character creation
  • 💬 Chat History: Persistent conversation tracking and management
  • 🎨 Modern UI: Clean, responsive interface with multiple themes built in
  • 🔍 Web search: Paste URL's into chat to ask question about them
  • 🖼️ Image Generation: Generate images from text prompts (currently only supports Replicate)
  • 📁 Image Gallery: View images that have been generated
  • 🖥️ Desktop Shortcut: Alt + N to open a new chat
  • 📝 Code Preview: Preview generated html, css and javascript in a separate window
  • 🔍 RAG Support: Upload documents and attach these to characters, who can then use them for their conversations
  • 🛠️ Tool Support: Recently added tool support - currently Note, Email and Web Search - plan to add more and allow user-defined tools

Roadmap

  • 🎤 Speech-to-Text (STT): Voice input capabilities
  • 📞 Voice Calls: Real-time voice conversations with AI
  • 📸 Vision Integration: Camera support for vision model capabilities
  • ⚙️ Custom Filters: Self-programmable filters and data processing pipes
  • 🗣️ Text-to-Speech (TTS): Natural voice output for AI responses (currently only supports ElevenLabs)
  • 🛠️ User defined Tools: Allow users to write their own tools for their characters.
  • 🛠️ Toolsmith: Create tool to create tools, so AI characters can build new tools for you from the chat
  • 🧑‍💼 Manager: Create tool to create/modify character, so AI can manage characters for you from the chat

Installation

Docker

To pull and run the Docker image:

docker pull ghcr.io/nordwestt/compass:latest
docker run --network host -p 80:80 ghcr.io/nordwestt/compass:latest

Then you can use compass by visiting "http://localhost/compass" in your browser. The proxy server will be available on port 3001.

Development

Get Started

It is recommended that you install Ollama and open up access, following the guide

  1. Install dependencies

    npm install
  2. Start the app

    nvm use 22.11.0
    npx expo start

You can run the app using:

Building app for Linux

First install the prerequisites: https://v2.tauri.app/start/prerequisites/

Then run:

nvm use 18.19.0
npm run build:linux

Building app for Docker

docker build -t my-app .

then you can run it with:

docker run -p 80:80 my-app

Make sure to update proxy in Settings to "http:///proxy"

Contributing

I don't currently accept contributions since I'm still working on the core functionality, but I welcome feedback and suggestions for now!

About

A provider-agnostic and privacy-respecting AI application, based on React Native and the AI SDK

Resources

Stars

10 stars

Watchers

2 watching

Forks

Releases

Packages

Used by

Contributors

Languages