Skip to main content
PromptQuorum
Home/Power Local LLM/Enchanted Review 2026: Native macOS/iOS Ollama Client
Easiest Desktop Apps

Enchanted Review 2026: Native macOS/iOS Ollama Client

Β·10 min readΒ·By Hans Kuepper Β· Founder of PromptQuorum, multi-model AI dispatch tool Β· PromptQuorum

**Enchanted is a free, open-source (Apache 2.0) native chat app for macOS, iOS, and visionOS that connects exclusively to a self-hosted Ollama server β€” it has no built-in model engine and no cloud-provider support.** It is built and maintained by Augustinas Malinauskas (GitHub: gluonfield, formerly published under the username AugustDev), has roughly 6,000 GitHub stars, and is distributed through the Apple App Store (listed as "Enchanted Developers Only") or by building the open-source Swift project yourself in Xcode. There is no paid tier, subscription, or in-app purchase of any kind.

Enchanted (github.com/gluonfield/enchanted) is an open-source, Apache-2.0-licensed chat app for macOS, iOS, and visionOS that connects to a self-hosted Ollama server. It does not run models itself β€” it is a native front end that talks to Ollama over HTTP, so you still need Ollama installed and a model pulled before Enchanted has anything to chat with. This review covers what the app actually does, its real project history (including a maintainer note that points at a newer, different project), how to download it, and where it fits next to other local chat clients.

Enchanted Review 2026: Native macOS/iOS Ollama Client

Key Takeaways

  • Enchanted only talks to a self-hosted Ollama server β€” it has no built-in inference engine and does not connect to OpenAI, Anthropic, or any other cloud provider
  • It is fully open source under the Apache 2.0 license, with the Swift source code public on GitHub
  • It runs natively on macOS, iOS, and visionOS (Apple Vision Pro) β€” there is no Windows, Linux, or Android build
  • It costs nothing: no subscription, no paid tier, no in-app purchase, on the App Store or when built from source
  • It has about 6,000 GitHub stars, 426 forks, and 114 open issues as of this review; the repository was created in December 2023
  • The author's own README now states that a separate, newer project called Jaz is "the new iteration of this project" β€” Enchanted is not archived, but it is worth knowing that before you invest time in it

πŸ“ In One Sentence

Enchanted is a free, open-source native macOS/iOS/visionOS chat app that connects only to a self-hosted Ollama server, with no cloud-model support, no paid tier, and roughly 6,000 GitHub stars.

πŸ’¬ In Plain Terms

Instead of chatting with Ollama through a web browser tab or a terminal command, Enchanted gives you a normal-looking iPhone, iPad, Mac, or Vision Pro app β€” with dark mode, voice input, image attachments, and conversation history β€” that just needs to know the address of your Ollama server.

πŸ“ŒNote: On the App Store, this app is listed under the name "Enchanted Developers Only," not simply "Enchanted." Search the App Store by that exact name, or use the direct link in the downloads section below, if a plain "Enchanted" search does not surface it.

What Is Enchanted?

**Enchanted is a native chat application for macOS, iOS, and visionOS that acts as a front end for a self-hosted Ollama server.** Per its own GitHub description, it is "an open source, Ollama compatible, elegant macOS/iOS/visionOS app for working with privately hosted models such as Llama 2, Mistral, Vicuna, Starling, and more." Enchanted itself never downloads, stores, or runs a model β€” it sends your prompts to whatever Ollama server address you configure and displays the streamed response.

  • Core function: a native SwiftUI chat interface that sends requests to an Ollama server's HTTP API and renders the response β€” nothing more, nothing less
  • Author: Augustinas Malinauskas, who develops under the GitHub username gluonfield (the repository was previously reachable under the username AugustDev, which GitHub now redirects to gluonfield)
  • License: Apache 2.0, meaning the source is free to read, modify, and redistribute, with attribution
  • "Native" here means it is a real macOS/iOS app built with Apple's SwiftUI framework β€” not a website wrapped in a browser shell β€” so it gets platform features like Spotlight-style command panels, native Markdown rendering, and offline-capable local storage of your chat history
  • What it does not do: it has no bundled model runtime (unlike LM Studio or GPT4All) and no cloud-provider connections (unlike Msty) β€” Ollama is the only supported backend

Enchanted's Project History and Version Milestones

Enchanted's first commits date to December 2023, with the GitHub repository created on December 13, 2023 and the App Store listing going live two days later, on December 15, 2023. The project shipped frequent releases through most of 2024, then had a roughly two-year gap in tagged GitHub releases before a burst of community-contributed features shipped in mid-2026.

  1. 1
    v1.1.0 β€” December 22, 2023: Voice recordings and image prompts
    Why it matters: Early releases added voice-to-text prompting and image attachments for multimodal models like LLaVA.
  2. 2
    v1.4.0 β€” February 12, 2024: macOS support
    Why it matters: Enchanted expanded from an iOS-only app to a native Mac build, adding a ⌘+N new-chat shortcut and Mac-specific settings.
  3. 3
    v1.4.4 β€” February 14, 2024: macOS Spotlight-style command panel
    Why it matters: Added the Ctrl+⌘+K command panel β€” a system-wide way to run a saved prompt against selected text without opening the main chat window.
  4. 4
    v1.5.5 "Completions" β€” March 11, 2024: Custom prompt templates
    Why it matters: Let users build reusable prompt templates that combine a saved instruction with text selected in any other Mac app.
  5. 5
    v1.6.7 β€” May 26, 2024: macOS text-to-speech and copy chat
    Why it matters: Added spoken "Read Aloud" output on macOS and a one-click way to copy an entire conversation.
  6. 6
    v1.7.0 β€” May 27, 2024: Apple Vision Pro support
    Why it matters: Enchanted became one of the earlier local-LLM chat clients with a dedicated visionOS build, plus access to Apple's higher-quality Enhanced/Premium system voices.
  7. 7
    v1.8.2 β€” June 15, 2026: Llama Vision, "think" tag support, and a color refresh
    Why it matters: After a long gap since v1.7.0, this release β€” built from several community pull requests β€” added support for Llama's vision-capable models, rendering for reasoning models' `<think>` tags, and interface color adjustments.

What Can You Do With Enchanted?

Enchanted's feature list, per its own README, is intentionally narrow: everything on it supports one job β€” chatting with your own Ollama server through a native app.

  • Text to speech ("Read Aloud") β€” has the app read a response aloud using the system voice, useful for listening to long answers hands-free
  • Voice prompts β€” dictate a prompt instead of typing it
  • Image attachments β€” attach an image to a prompt for multimodal models such as LLaVA or Llama's vision-capable variants (added in v1.8.2)
  • Custom system prompts β€” set a system prompt that applies to a specific conversation
  • Edit and resend β€” edit a previous message and resend it, optionally to a different model than the one that originally answered it
  • Markdown rendering β€” tables, lists, and code blocks (with a copy button) render properly instead of showing raw Markdown syntax
  • On-device conversation history β€” chat history is stored locally on your device, not on a company server, since there is no company server involved at all
  • macOS command panel (Ctrl+⌘+K) β€” a Spotlight-style panel for running a saved custom prompt template against text selected in any other Mac application, without switching to the main Enchanted window
  • Reasoning-model support β€” renders <think> tags from reasoning-style models as a distinct block rather than raw text (added in v1.8.2)

Usage Examples: Three Ways to Use Enchanted

These three workflows come directly from Enchanted's documented setup instructions and features β€” not hypothetical scenarios.

Enchanted Pricing: Free and Open Source

Enchanted costs $0. There is no subscription, no paid tier, and no in-app purchase β€” on the App Store or in the source code. This is a straightforward case, not a freemium product with a limited free layer: every feature described in this review, including image attachments, voice prompts, and the macOS command panel, is available at no cost.

  • App Store price: Free (confirmed via the App Store listing's pricing field)
  • No subscription tier of any kind exists for this app, unlike several other apps covered in this cluster
  • Building from source costs nothing beyond an Apple Developer account, which Apple requires to sign and install any app you compile yourself β€” a $99/year Apple Developer Program membership if you want to install a self-built copy on a physical device long-term, or a free personal-team signing option that requires periodic re-signing
  • The only real "cost" is indirect: you still need to run Ollama and download models yourself, and those use your own machine's compute and storage, not a paid API

Enchanted vs. Msty: Which One Do You Want?

**Enchanted and Msty solve a similar surface problem β€” a desktop chat interface for local models β€” with almost opposite designs.** Enchanted is a narrow, open-source, Ollama-only native app for Apple platforms. Msty is a closed-source, cross-platform app that connects to multiple local runtimes and cloud providers side by side, with a paid tier for extra features.

Aspect
Enchanted
Msty
Source codeOpen source (Apache 2.0)Closed source
PlatformsmacOS, iOS, visionOS onlymacOS, Windows, Linux
Model backendsOllama onlyOllama, LM Studio, llama.cpp, MLX + cloud APIs
Cloud model supportNoneYes, via provider API keys
Document RAG searchNot built inYes (Knowledge Stacks)
PriceFree, no paid tierFree tier + paid Aurum tier

Pick Enchanted if you want a minimal, native, open-source app that only talks to your own Ollama server and nothing else. Pick Msty if you also want cloud-provider models, document search over your own files, or a Windows/Linux build β€” Enchanted does not offer any of those.

Who Should Use Enchanted?

Enchanted fits a specific, narrow use case well; it is a poor fit outside that case.

Enchanted vs. Other Local Chat Apps

Enchanted is one of several chat front ends built specifically around Ollama, and one of the few native Apple-platform options among them. Here is how it compares β€” see the Local LLM Software Directory for the full catalog.

  • Msty** β€” a closed-source, cross-platform (macOS/Windows/Linux) app that adds cloud-model support and document RAG search on top of local-model chat; see the full Enchanted vs. Msty comparison above.
  • Msty Go** β€” a separate autonomous-agent product from the Msty team, not a chat interface at all; not a direct alternative to Enchanted.
  • LM Studio, Jan, and GPT4All** β€” three cross-platform local chat clients that, unlike Enchanted, bundle their own model runtime rather than requiring a separate Ollama installation.
  • Jan** β€” an open-source, local-first chat app with a bundled inference engine, running on macOS, Windows, and Linux, unlike Enchanted's Ollama-only, Apple-only scope.
  • GPT4All** β€” an open-source local chat client from Nomic AI with its own bundled runtime, available on Windows, macOS, and Linux.
  • Open WebUI** β€” a self-hosted, browser-based chat interface commonly paired with Ollama; unlike Enchanted, it runs as a web app accessible from any device with a browser rather than a native mobile/desktop client.
  • BoltAI** β€” a macOS/Windows AI assistant with deep OS-level integration (system-wide shortcuts, app integrations) and both local and cloud model support, closer to Msty's scope than Enchanted's.
  • Jaz** β€” the same author's newer, separate project: a "personal AI agent" platform (always-on agents, scheduled loops, memory, git control) rather than a simple chat client; not a like-for-like replacement for Enchanted's use case.

Common Mistakes When Evaluating Enchanted

Most confusion about Enchanted comes from assuming it does more than a minimal Ollama front end, or missing its App Store listing name.

Frequently Asked Questions

What is Enchanted?

Enchanted is a free, open-source native chat app for macOS, iOS, and visionOS that connects to a self-hosted Ollama server. It has no built-in model engine β€” it is a front end only, per its GitHub repository.

Is Enchanted free?

Yes, fully. There is no subscription, paid tier, or in-app purchase, whether you install it from the App Store or build it from source. The only ongoing cost is whatever your own hardware and Ollama setup already cost you.

Does Enchanted work with anything besides Ollama?

No. Per its own documentation, Enchanted is "Ollama compatible" and connects exclusively to a self-hosted Ollama server. It does not connect to LM Studio, llama.cpp directly, or any cloud provider such as OpenAI or Anthropic.

What platforms does Enchanted support?

macOS, iOS, and visionOS (Apple Vision Pro). There is no Windows, Linux, or Android build in any form, on the App Store or from source.

Is Enchanted open source?

Yes. The full Swift source code is published on GitHub under the Apache 2.0 license, meaning it can be freely read, modified, and redistributed with attribution.

Who makes Enchanted?

Augustinas Malinauskas, who develops under the GitHub username gluonfield (the project was previously published under the username AugustDev, which now redirects to gluonfield).

What is the App Store listing called?

The app is listed as "Enchanted Developers Only," not simply "Enchanted." Search for that exact name, or use the direct App Store link, if a plain search does not find it.

Can I use Enchanted on my iPhone with Ollama running on my Mac?

Yes β€” this is one of the two setups documented in the project's own README. Expose your Mac's Ollama server with a tool like ngrok, then enter the resulting URL as the server endpoint in Enchanted on your iPhone.

Is Enchanted still maintained?

The repository is not archived and its most recent code push was July 2026, but there was a roughly two-year gap between its v1.7.0 (May 2024) and v1.8.2 (June 2026) tagged releases. Its own README also now points to a separate, newer project called Jaz as "the new iteration of this project," so check that project's scope before assuming Enchanted will keep receiving frequent updates.

Does Enchanted support image or vision prompts?

Yes. It supports image attachments for multimodal models such as LLaVA, and added support for Llama's vision-capable models in the v1.8.2 release (June 2026).

Sources

← Back to Power Local LLM