explaingit

gluonfield/enchanted

★ 5,929SwiftAudience · generalComplexity · 2/5Setup · moderate

TLDR

Enchanted is a polished iPhone, iPad, and Mac chat app that connects to AI language models running on your own computer via Ollama, keeping all conversations private and working offline.

Mindmap

mindmap
  root((enchanted))
    What It Does
      Chat with local AI
      No cloud required
      Works offline
    Supported Platforms
      iPhone and iPad
      Mac
      Apple Vision Pro
    Key Features
      Voice input
      Image attachments
      Custom system prompts
    Setup
      Install Ollama
      Download AI model
      Point app at server
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Chat with AI models like Llama 2 or Mistral from your iPhone without sending any data to an external cloud service.

USE CASE 2

Set up a private AI assistant on your Mac and access it from iPhone, iPad, and Apple Vision Pro using Enchanted.

USE CASE 3

Use voice input and image attachments to interact with a locally-hosted AI model through a native Apple app.

Tech stack

SwiftOllama

Getting it running

Difficulty · moderate Time to first run · 30min

Requires Ollama installed on a Mac or Linux computer with at least one AI model downloaded before the app can connect.

In plain English

Enchanted is an open-source chat app for iPhone, iPad, Mac, and Apple Vision Pro that connects to AI language models you run yourself. Instead of sending your messages to a cloud service like ChatGPT, it talks to a local server on your own machine or network. This keeps conversations private and lets you use the app without an internet connection once everything is set up. The app works with Ollama, a free tool that lets you download and run AI language models locally on a Mac or Linux computer. Models you can use include Llama 2, Mistral, Vicuna, Starling, and others. You install Ollama on a computer, download one or more models through it, and then point Enchanted at that server. If your Ollama server is only accessible on your local network, the README describes using ngrok, a tunneling tool that creates a temporary public URL so your phone can reach your computer. The interface is designed to feel like a polished chat app. Features include conversation history stored on your device, markdown rendering for code blocks and tables, image attachments for multimodal prompts, voice input, text-to-speech readback, custom system prompts that apply to every conversation, and a macOS quick-access panel. All features work offline once the server is reachable. The app is available on the Apple App Store. Source code is written in Swift. The README is primarily a visual showcase with setup instructions for two scenarios: running Ollama with public access already configured, or running it locally and exposing it through ngrok. No license is specified in the README.

Copy-paste prompts

Prompt 1
I have Ollama running on my Mac with Llama 2 downloaded. How do I connect the Enchanted iPhone app to it and start chatting?
Prompt 2
Show me how to use ngrok to expose my local Ollama server so the Enchanted app on my iPhone can reach it from outside my home network.
Prompt 3
How do I set a custom system prompt in Enchanted so every conversation with my local AI model follows specific instructions or a persona?
Open on GitHub → Explain another repo

← gluonfield on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.