# NeutralLM — Run language models locally

> Run language models locally on your device for private experimentation. (iPhone/iPad app by Neutral Apps LLC.)

- Source: https://appshunter.io/ios/app/neutrallm/id6752721725 (this page in markdown: same URL + `.md`)
- Developer: [Neutral Apps LLC](https://appshunter.io/developer/1817554179)
- Category: Developer Tools, Reference
- Price: Free
- Rating: 4.00/5 from 1 App Store ratings · 2 written reviews indexed
- Age rating: 12+
- Requires: iOS 18.5 · 33 MB
- Languages: American English
- Released: 2025-09-22
- Data updated: 2026-08-26
- Monetization: free
- User reviews in markdown: https://appshunter.io/ios/app/neutrallm/id6752721725/reviews.md

## What is NeutralLM?

Discover, download, and run language models directly on your device.

Neutral LM makes it simple to explore open-source models and run them locally on your iPhone, iPad, or Mac — no cloud required. Keep your data on your device while experimenting with different models and workflows.

Key Features
-Browse and download third-party models from online sources
-Run models fully offline once installed — your prompts and outputs stay on your device
-Flexible setup so you can choose which models to install and use
-Privacy-first design: Neutral LM does not collect your prompts, outputs, or files

Why Neutral LM?
Neutral LM is built for developers, researchers, and curious users who want more control over language model experimentation. By running models locally, you stay in control of your data and can work without relying on external servers.

Disclaimer: Neutral LM lets you discover and download third-party language models. Once installed, models run locally on your device — your prompts and outputs are not sent to us. You are solely responsible for the models you choose, the content generated, and how you use it. Do not use Neutral LM for unlawful, harmful, or exploitative purposes.

## Key features

- Discover and download third-party models
- Run models fully offline
- Flexible model setup
- Privacy-first design
- Local data storage

## Recent user reviews (2 of 2)

All indexed reviews: https://appshunter.io/ios/app/neutrallm/id6752721725/reviews.md

### 4/5 — No supports LaTeX / Math rendering

*2026-04-11*

:::writing{variant=“standard” id=“40218”}
Please add native LaTeX math rendering support for Markdown output.

Requirements:
	•	inline: $...$
	•	block: $$...$$
	•	Prefer native iOS rendering with SwiftMath
	•	Avoid WebView if possible
	•	Fallback to Unicode math symbols
	•	Support GGUF / MLX local models
:::

### 1/5 — Frequent crashes

*2026-03-12*

Crashes every time it’s opened after loading a model.

## Frequently asked questions about NeutralLM

### What is NeutralLM?

NeutralLM allows you to discover, download, and run language models directly on your iPhone, iPad, or Mac. It enables local model experimentation without requiring cloud access.

### Does NeutralLM require an internet connection to run models?

No, once models are downloaded and installed through NeutralLM, they can be run fully offline. Your prompts and outputs will remain on your device.

### Is NeutralLM free to use?

Yes, NeutralLM is available for free. There are no costs associated with downloading or using the application itself.

### What devices does NeutralLM support?

NeutralLM is compatible with Mac, iPhone, and iPad devices, allowing for flexible use across Apple's ecosystem.

### How does NeutralLM handle user data and privacy?

NeutralLM is designed with a privacy-first approach. It does not collect your prompts, outputs, or files, ensuring your data stays on your device.

### How often is NeutralLM updated?

The latest version of NeutralLM is 0.16, and it was last updated on April 20, 2026. This indicates a relatively recent update for the application.

### What is the age rating for NeutralLM?

NeutralLM has an age rating of 12+, suggesting it is suitable for users aged 12 and above.

### Can I use NeutralLM for commercial purposes?

The developer's disclaimer states users are solely responsible for the models they choose, the content generated, and how they use it. It advises against unlawful, harmful, or exploitative purposes, implying responsible use is expected.

## Version history (last 5 releases)

### 0.16 — 2026-04-20

Rich Text & Math                                                                                                         
- Assistant replies now render formatting and equations. Markdown (headings, lists, code blocks, block quotes) and LaTeX math — both inline and display — show as intended instead of raw source. Great for STEM, code-heavy, and long-form responses.
- Can be turned off in Settings if you prefer plain text.

### 0.15 — 2026-04-03

- support for gemma 4 and other new models
- now using llama.cpp and mlx directly
- bug fixes
- improved compatibility filtering

### 0.14 — 2026-01-07

Improved model compatibility and updates

### 0.13 — 2025-10-27

-auto generate chat titles
-faster load time
-device compatibility by file

### 0.12 — 2025-10-11

-MLX support with vision
-GGUF vision experimental support
-fixed issue where filter control hidden when no results
-fixed sidebar token count issue and delete/rename issue
-package upgrades
-improved stop generation / error handling
-macOS fixed local model context menu

## More apps by Neutral Apps LLC

- [Tilestacks](https://appshunter.io/ios/app/tilestacks/id6746613122)

All apps by Neutral Apps LLC: https://appshunter.io/developer/1817554179

---

*Data collected daily from the US App Store and indexed by [AppsHunter](https://appshunter.io/). User reviews are verbatim App Store reviews. Ratings, prices and chart positions refresh continuously; this snapshot is from 2026-08-26.*