Enchanted is an open-source, Ollama-compatible elegant macOS/iOS/visionOS app for conversing with private self-hosted models (such as Llama 2, Mistral, Vicuna, Starling, etc.). It is essentially a ChatGPT-style interface for connecting to private models, with the goal of providing an unfiltered, secure, private, and multimodal experience across the entire iOS ecosystem (macOS, iOS, Watch, Vision Pro). The project has entered a new phase, and Jaz is its new iteration.
Main capabilities include: conversation history; text-to-speech (read aloud); including conversation history in API calls; dark/light mode; conversation history stored on the device; Markdown support (can display tables/lists/code blocks well); prompt support for image attachments; can specify a system prompt for each conversation; can edit message content or submit messages with a different model; can delete a single conversation or delete all conversations; macOS Spotlight panel (Ctrl + ⌘ + K); all features can be used offline.
How to use: Enchanted requires Ollama v0.1.14 or later. Scenario one: if running the Ollama server with public access - download the Enchanted app from the App Store, specify the server endpoint in the app settings, and start prompting. Scenario two: if running Ollama on your own computer - start the Ollama server and download the required models; install ngrok and forward the Ollama server to be publicly accessible via ngrok http 11434 --host-header="localhost:11434"; copy the "Forwarding" URL in the form https://b377-82-132-216-51.ngrok-free.app , and the Ollama server API is accessible through that temporary URL; specify the server endpoint in the app settings and start prompting.
Use cases: users who want to privately use self-hosted large models on Apple devices can use Ollama to connect to private models, converse locally or over the network, and take advantage of capabilities such as image attachments, system prompts, read aloud, and Markdown display.
Limitations: you must run the Ollama server yourself to use the app; Ollama v0.1.14 or later is required; if Ollama runs on a personal computer and needs remote access, port forwarding via tools such as ngrok is required. Other information such as fees, pricing, platform details, or performance is not currently provided.
