LlamaChat Local

Run AI models 100% on-device. Download Llama, Phi, Gemma, and Mistral models, then chat privately with no internet required. Your data never leaves your device.

Chat with AI models privately and entirely on your device. LlamaChat Local lets you download and run open-source large language models like Llama, Phi, Gemma, and Mistral — all powered by llama.cpp with Metal GPU acceleration for fast, efficient inference.

Your conversations never leave your iPhone. All AI processing happens 100% on-device with no network calls after downloading a model. Conversations are saved locally so you can pick up right where you left off.

Features:

• Browse and download GGUF models directly from HuggingFace
• Run Llama 3.2 1B, Llama 3.2 3B, Phi-3.5 Mini, Gemma 2 2B, and more
• Real-time token streaming with tokens/second display
• Adjustable temperature, context window (512-8K), and system prompt
• Track download progress and manage local model storage
• Dark mode interface with clean, modern design
• No account required — just download a model and start chatting

LlamaChat Local is perfect for anyone who values privacy, wants to experiment with open-source AI models, or needs an offline AI assistant.