Enchanted Review, Pricing & Alternatives (2026)

Enchanted is an open-source Swift and SwiftUI client for chatting with privately hosted models through Ollama. Licensed under Apache 2.0, it supports macOS, iOS, and visionOS, with documented compatibility including Llama, Mistral, Vicuna, Starling, and multimodal models. The app provides Markdown chat, supported image prompts, custom completions, Apple voices, and local conversation history, but no inference engine, cloud access, subscription, model downloads, or marketplace. Users must provide an Ollama server, model files, compatible hardware, storage, and a reachable endpoint, while remote exposure requires their own security controls. Enchanted remains the named Apple client, although its README identifies Jaz as the project's new iteration and newer agent work.

Using Enchanted requires Ollama version 0.1.14 or later, a downloaded model, and Apple hardware for its native client. The application connects to an Ollama server running on hardware you control, whether locally or on another reachable machine, then sends prompts through Ollama’s HTTP API. Setup involves entering the server endpoint, and remote connections may require a public endpoint or an ngrok tunnel, leaving authentication, transport security, firewall rules, and access control to the user. Enchanted offers native macOS, iOS, and visionOS chat with Markdown rendering, local conversation history, multimodal prompts for compatible models, custom system prompts, message editing, model selection, and Apple voice features. It provides no model runtime, downloads, hosted inference, document retrieval, or marketplace.

Enchanted is a native Swift and SwiftUI Apple client for Ollama's HTTP API, not an independent model runtime. Its Apache 2.0 repository describes an open-source application for macOS, iOS, and visionOS that connects to models running on hardware the user controls. The app requires separate Ollama infrastructure, model files, compute, and storage, and offers no hosted inference, subscription, paid tier, document retrieval, or model marketplace. Release v1.7.0 added Vision Pro and Apple voice support, while v1.8.2, published June 15, 2026, added visionOS work, Llama Vision support, tag handling, color adjustments, and conversation fixes. The maintainer identifies Jaz as the new iteration, with agents, memory, boards, Git control, and a client/server architecture.

Enchanted’s application source is available under Apache 2.0, and the project lists no subscription, paid plan, in-app purchase, or published usage quota. The app itself therefore costs $0, but users must supply Apple hardware for the native macOS, iOS, or visionOS client, plus an Ollama installation, model storage, compatible compute, electricity, and any remote networking. Ollama can run on the same device or another reachable machine, although exposing it publicly requires users to manage authentication, transport security, firewall rules, and access control. Apache 2.0 covers Enchanted’s source, not automatically Ollama, model weights, operating systems, hosted infrastructure, or other dependencies. Commercial use consequently requires reviewing each component’s separate terms.

Enchanted is narrower than Open WebUI, LM Studio, and Jan. Open WebUI provides a browser-based self-hosted interface with broader server workflows, LM Studio bundles local model management on desktop, and Jan offers an open-source cross-platform client with a local runtime path. Enchanted is instead a free, Apache 2.0 native macOS, iOS, and visionOS interface that connects directly to an Ollama server and supports Markdown, multimodal prompts, custom completions, Apple voices, and local history. It supplies no runtime, model downloads, cloud providers, hosted account, document retrieval, marketplace, or agent automation. Its Apple-only distribution also requires users to manage Ollama, models, hardware, storage, and networking. The README identifies Jaz as the newer direction. Pick Enchanted for simple private Ollama chat on Apple devices.