# Private LLM

> Private LLM is a one-time purchase Local AI chat app for iPhone, iPad, and Mac. It runs open-source language models such as DeepSeek R1 Distill, Llama 3.3 70B, Qwen3 4B, Phi 4, and Gemma 3 entirely on-device, with OmniQuant and GPTQ quantization tuned per model. No internet after model download, no account, no subscription, and the app collects no personal data.

Private LLM is built by Numen Technologies Limited, registered in Ireland (CRO 677823), a bootstrapped team of two engineers in the EU. The inference engine is proprietary; the models it runs are open-source. OmniQuant and GPTQ are external research methods that Private LLM ships and tunes per model. The app is distributed through the Apple App Store, with Family Sharing for up to six people.

Hardware guidance: Apple Silicon Macs with 48GB or more RAM can run Llama 3.3 70B. iPhones and iPads with 4GB RAM run models up to 3B parameters well. iPhone 15 Pro and other 8GB devices handle 7B and 8B models such as Llama 3.1 8B and Qwen 2.5 7B. No iPhone runs a 70B model.

## Key pages

- [Home](https://privatellm.app/en): Product overview, features, and supported models
- [FAQ](https://privatellm.app/en/faq): Technical answers, hardware requirements, refunds, Family Sharing, and comparisons with Ollama and LM Studio
- [Release Notes](https://privatellm.app/en/release-notes): Version history for the iOS and macOS apps
- [Reviews](https://privatellm.app/reviews): App Store review archive for iOS and macOS
- [Community Shortcuts](https://privatellm.app/en/community-shortcuts): User-built Siri and Apple Shortcuts workflows that call Private LLM
- [Blog](https://privatellm.app/blog): Model announcements, guides, and roundups

## Models

- [Models Catalogue](https://privatellm.app/models): Every model Private LLM runs, filterable by device, RAM, and use case
- [Model Families](https://privatellm.app/models/family): Index of supported model families
- [Best Models for iPhone](https://privatellm.app/models/best-for-iphone): Models sized for iPhone and iPad RAM tiers
- [Best Models for Mac](https://privatellm.app/models/best-for-mac): Models for Apple Silicon Macs, up to 70B
- [Coding Models](https://privatellm.app/models/coding): Local models for code generation and review
- [Reasoning Models](https://privatellm.app/models/reasoning): Models with step-by-step reasoning, such as DeepSeek R1 Distill
- [Uncensored Models](https://privatellm.app/models/uncensored): Abliterated and uncensored model variants
- [Roleplay Models](https://privatellm.app/models/roleplay): Models tuned for role-play and story writing
- [Multilingual Models](https://privatellm.app/models/multilingual): Models for languages other than English
- [Biomedical Models](https://privatellm.app/models/biomedical): Medical and biomedical domain models
- [Survival Models](https://privatellm.app/models/survival): Models for offline use when no network is available
- [DeepSeek R1 Distill](https://privatellm.app/models/family/deepseek-r1-distill): DeepSeek R1 Distill reasoning models
- [Llama 3.3 70B](https://privatellm.app/models/family/llama-3-3-70b): Meta's 70B model for Macs with 48GB or more RAM
- [Qwen3 4B](https://privatellm.app/models/family/qwen3-4b): Qwen3 4B and its fine-tunes
- [Phi 4](https://privatellm.app/models/family/phi-4-14b): Microsoft Phi 4 14B
- [Gemma 3 1B](https://privatellm.app/models/family/gemma-3-1b): Google Gemma 3 1B

## Comparisons

- [Local AI Chatbots Compared](https://privatellm.app/compare-local-ai-chatbots): Private LLM against other local AI chat apps
- [Private LLM vs Ollama](https://privatellm.app/compare/ollama-vs-private-llm): Quantization quality and iOS support differences
- [Private LLM vs LM Studio](https://privatellm.app/compare/lm-studio-vs-private-llm): Native app compared with LM Studio's desktop app
- [Private LLM vs LLM Farm](https://privatellm.app/compare/llm-farm-vs-private-llm): Mac and iOS coverage and UX comparison
- [Private LLM vs PocketPal AI](https://privatellm.app/compare/pocketpal-ai-vs-private-llm): Feature and model support comparison
- [Private LLM vs Apollo AI](https://privatellm.app/compare/apollo-ai-vs-private-llm): Feature and model support comparison

## Guides

- [Run Local GPT on iOS: Complete Guide](https://privatellm.app/blog/run-local-gpt-on-ios-complete-guide): Setup, sampling settings, system prompts, model sizing, and integrations
- [Understanding System Prompts](https://privatellm.app/blog/run-local-gpt-on-ios-complete-guide#understanding-system-prompts): Writing system prompts (section)
- [Understanding Sampling Settings](https://privatellm.app/blog/run-local-gpt-on-ios-complete-guide#understanding-sampling-settings): Temperature and Top-P tuning (section)
- [Best AI Coding Models to Run Locally](https://privatellm.app/blog/best-ai-coding-models-run-locally-iphone-ipad-mac): Coding model roundup for iPhone, iPad, and Mac
- [Best Uncensored LLMs to Run Locally](https://privatellm.app/blog/best-uncensored-ai-chat-run-locally-iphone-ipad-mac-2026): Uncensored model roundup
- [Heretic vs Abliterated](https://privatellm.app/blog/heretic-vs-abliterated-uncensored-llm-comparison): Comparison of two uncensoring methods
- [Survival AI Offline](https://privatellm.app/blog/run-survival-ai-offline-iphone-ipad-mac-private-llm): Using local AI with no network connection
- [DeepSeek R1 Distill in Private LLM](https://privatellm.app/blog/deepseek-r1-distill-now-available-private-llm-ios-macos): DeepSeek R1 Distill on iOS and macOS
- [Run Phi 4 Locally on Mac](https://privatellm.app/blog/run-phi-4-locally-mac-private-llm): Phi 4 on Apple Silicon Macs
- [Llama 3.3 70B on Mac](https://privatellm.app/blog/llama-3-3-70b-available-locally-private-llm-macos): Running Llama 3.3 70B locally on macOS

## Policies

- [Privacy Policy](https://privatellm.app/privacy-policy): What the app and the website do and do not collect
- [Terms of Use](https://privatellm.app/terms): Site terms, acceptable use, model licenses, and governing law
- [Press Kit](https://privatellm.app/press-kit): Brand assets and media resources

## Optional

- [Summarize Articles Shortcut](https://privatellm.app/shortcuts/summarize-articles-using-local-ai-with-private-llm-and-apple-shortcuts): Example Apple Shortcut using Private LLM
- [Chat with Screenshot Shortcut](https://privatellm.app/shortcuts/chat-with-screenshot): Example Apple Shortcut using Private LLM