Skip to main content
PromptQuorumBuilt for humans. Structured for AI.
Home/Power Local LLM
Mobile & Edge LLMs

LM Mini Review: iPhone Chat App for On-Device and Home-Server AI

·6 min readBy Hans Kuepper · Founder of PromptQuorum · Discovery engine for open-weight & open-source AI

LM Mini is a closed-source chat app for iPhone and iPad that combines on-device models with a connection to LM Studio, Ollama, Jan AI or oMLX on your own computer. Its privacy policy dates from August 2025 and describes only the LM Studio connection, so check what the app does today before sending sensitive chats.

Key Takeaways

  • Developer: Sultan Haider, who published the first release on 20 September 2025; version 1.10.4 followed on 19 October 2026, and the listing shows 25 releases since July 2026.
  • Reach: the US App Store shows 4.5 stars from 152 ratings, a modest but real track record.
  • Platforms: iOS and iPadOS 16.0 or later and macOS 14.0 or later, plus an Apple Watch app; no Android build was found.
  • Website: the listing links no developer website, and the only public developer page found is the privacy policy.

📍 In One Sentence

LM Mini is a closed-source iPhone, iPad and Mac chat app that can run models on the device, connect to a model server on your own computer, or reach that computer remotely.

💬 In Plain Terms

You install it from the App Store and pick a mode: download a model to the phone, point the app at LM Studio, Ollama, Jan AI or oMLX on your network, or use the developer's remote service. Cloud providers such as OpenAI are optional extras.

📌Note: This review is based on the App Store listing, Apple's lookup data and the developer's privacy policy, checked on 24 October 2026. It involved no hands-on use, network inspection or benchmarking.

Read in:
🇺🇸en🇩🇪de🇫🇷fr🇯🇵ja🇨🇳zh🇪🇸es🇧🇷pt🇸🇦ar🇰🇷ko

LM Mini is a free chat app for iPhone, iPad and Mac from the developer Sultan Haider. According to its App Store listing it can run models on the device, connect to LM Studio, Ollama, Jan AI or oMLX on your own computer, or reach that computer remotely. This review reflects the store listing and the developer's privacy policy, as checked on 24 October 2026.

LM Mini Review: iPhone Chat App for On-Device and Home-Server AI

What Is LM Mini?

LM Mini is a chat client for local and remote language models on Apple devices. The App Store subtitle is "AI Chat for Ollama & LM Studio", and the listing describes one app for "phone-side models and your desktop local setup".

The developer is listed as Sultan Haider; the App Store shows no company name or website, and the listing names no affiliation with LM Studio, Ollama, Jan AI or oMLX.

Where Your Chats Run

LM Mini is hybrid, not purely local: the listing describes four paths, and where a chat runs depends on the mode you pick.

  • On the phone. You download a model and chat with no internet. The listing says phone models are smaller than hosted ones and that speed depends on the device. Apple Intelligence chat uses the model built into the device and needs no download.
  • On your own computer. The app connects to LM Studio, Ollama, Jan AI or oMLX over Wi-Fi or USB; the model then runs on that computer, not on the phone.
  • Through LM Mini Connect. A remote mode for reaching your computer away from home, "encrypted with your key". How the traffic is routed is not described.
  • In the cloud, if you add it. Optional API providers such as OpenAI, OpenRouter, Mistral, DeepSeek and Google Gemini receive your requests, and web search through SearXNG sends queries to a search server. Optional encrypted cloud backups also exist.

Get It

Platform
Where to get it
iPhone and iPadApp Store (free download, iOS or iPadOS 16.0 or later)
MacSame App Store listing (macOS 14.0 or later); it does not say whether this is a native Mac build
WebsiteNo official website found; the App Store listing links none
Privacy policyLM Mini privacy policy (last updated 16 August 2025)

Version reviewed: 1.10.4, released on 19 October 2026. Connected mode needs LM Studio, Ollama, Jan AI or oMLX running on a computer. The 1.10.4 notes add Apple Intelligence chat, a simpler setup that shows servers found on Wi-Fi, a choice between bubble and full-width chat style, and an improved Pro page.

Pricing

The app is a free download with a paid tier sold as in-app purchases: $4.99 per month, $29.99 per year or $119.99 once for lifetime access, as shown on the US App Store.

The listing does not say which features need a purchase. Prices differ by storefront.

Features Listed by the Developer

Everything below comes from the App Store description and release notes; none of it was tested.

  • Chat. Conversation history with folders and context-usage tracking, personas, group chat with up to five models, memory, on-device semantic search across chats and background generation.
  • Models and servers. Download and run on-device models, connect to custom servers with custom headers, manage and download models on LM Studio from the phone, and import chats from LM Studio.
  • Voice and images. Voice chat with Kokoro, and image generation through ComfyUI and Automatic1111.
  • Tools and search. Web search through SearXNG and MCP tool calling; since version 1.10.3, LM Studio MCP tools are also available from Siri and Shortcuts.
  • Apple devices. An Apple Watch app since version 1.10.0, and Apple Intelligence chat since version 1.10.4.

Privacy Policy and Data Flow

The privacy policy says the app does not collect, store or transmit personal data to external servers or third parties, and that all processing happens between your device and your local LM Studio server. It lists chat history, settings and server connection details as stored on the device and names no analytics or advertising services.

The policy was last updated on 16 August 2025, five weeks before the first release, and it does not mention on-device models, Ollama, Jan AI, oMLX, LM Mini Connect, cloud API providers or encrypted cloud backups. The App Store privacy label says "Data Not Collected" and notes that Apple has not verified it; the description adds that requests to optional cloud providers go to the provider you pick.

📌Note: The app is closed source, so none of these statements can be checked in code, and PromptQuorum has not inspected network traffic. Anyone using cloud APIs, LM Mini Connect or cloud backups should confirm in the app which data leaves the device.

Who Should Use It

  • iPhone and iPad owners who already run LM Studio, Ollama, Jan AI or oMLX. One app covers the home server and, when away from it, the phone-side models.
  • People who want to try on-device models without a server. Download a model, or use Apple Intelligence on supported devices.
  • Users of ComfyUI, Automatic1111 or MCP tools who want them inside a phone chat client.

What We Could Not Verify

  • Hands-on behavior. This review did not time answers, test offline use or measure battery and storage.
  • The on-device runtime and models. The listing names no inference engine, model list or storage need for on-device mode.
  • How LM Mini Connect works. The listing says it is encrypted with a key you control; it does not describe the route the traffic takes.
  • What the paid tier unlocks. The listing does not separate free from paid features.
  • Setup friction. Two public App Store reviews describe problems: one reports that iOS never asked for local network access (March 2026), the other says a desktop companion app is required, which the developer rejects (July 2026).
  • Source code and Mac build. The app is closed source with no public repository found, and the listing does not say whether the Mac version is native or the iPad app running on a Mac.
  • Languages. The listing shows English as the only language, although the release notes mention German and Russian voices.

Competitors and Alternatives

Enchanted

Platforms:
iOS, macOS
Price / license:
Free / Apache 2.0
Key difference:
Ollama-only client with no models of its own
Articles about Enchanted (4)

Also mentioned in:

+7 more not shown

AnythingLLM Mobile

Platforms:
iOS, Android
Price / license:
Free / MIT
Key difference:
Remote access to your local AnythingLLM workspace
Articles about AnythingLLM Mobile (2)

Also mentioned in:

Competitor details change often; confirm each app's current price, license and platforms on its own listing.

Compare in directory

Frequently Asked Questions

Does LM Mini run on Android?

No Android build was found; the listing covers iPhone, iPad, Mac, Apple Watch and Apple Vision devices only.

What is Apple Intelligence chat in LM Mini?

Added in version 1.10.4, it chats with the model built into the device, with no download, on supported iOS 26 and macOS 26 devices. It is separate from downloaded models and from your own server.

Where does LM Mini store API keys?

The version 1.10.3 release notes say API keys, tokens and server headers are now kept in the Keychain / Keystore instead of plain settings.

Verdict

LM Mini is a feature-rich, actively updated iPhone and iPad client that treats "local" as a spectrum: models on the phone, models on your own computer, a remote route to it and optional cloud APIs. Its limits are a closed-source codebase, a privacy policy that describes only the LM Studio connection and an unexplained remote service. It suits readers who already run LM Studio, Ollama or Jan AI; anyone who needs a fully on-device, open-source app should compare Private Mind or PocketPal AI.

Sources

← Back to Power Local LLM

MCP