MyLLM turns your local LLM server into a powerful AI agent on your iPhone or iPad β with 60+ built-in tools, voice control, and automation that no cloud-based assistant can match. Now with a built-in server terminal, on-device offline inference, and smart mode switching. YOUR AI. YOUR NETWORK. YOUR PRIVACY. Unlike ChatGPT, Siri, or Google Assistant, MyLLM connects to an LLM running on YOUR computer. Your conversations, files, and personal data never leave your home network. Zero cloud. Zero subscriptions. Zero data harvesting. Works with Ollama, LM Studio, or any OpenAI-compatible local server. Supports Qwen 2.5, Llama 3.2, Mistral, DeepSeek, Phi-3, Gemma, and any model you can run. β SERVER TERMINAL: YOUR SERVER IN YOUR POCKET β The first iOS app with a full built-in terminal for managing your Ollama server. No more switching to your desktop to manage models β do it all from your iPhone whether your server is on Windows or Mac: $ pull qwen2.5:7b Download models with real-time streaming progress $ list View all installed models β name, size, family, quantization $ ps See which models are loaded in memory and their VRAM usage $ show mistral:7b Inspect any model: parameters, template, license, format $ rm old-model Free up disk space by removing models you no longer need $ cp model new-name Duplicate or rename models instantly Three powerful views in one screen: 1. Terminal β A real command-line interface with a green $ prompt, monospaced output, and quick-tap shortcut buttons for common commands. Type commands or tap to execute. 2. Model Browser β A visual list of every model on your server. See size, quantization level, model family, and last modified date at a glance. Swipe left to delete, swipe right for detailed info. 3. Running Monitor β Live view of which models are currently loaded in your server's memory, how much VRAM they're using, and when they'll be unloaded. Pull to refresh. Smart server detection automatically identifies whether you're connected to Ollama, LM Studio, or a custom OpenAI-compatible server β and adjusts available commands accordingly. Ollama-only commands (pull, rm, show, cp, ps) are clearly gated so you never send an incompatible request to the wrong server type. Whether your server runs on Mac, Windows, or Linux β manage it from the couch, from bed, or from another room. No SSH, no terminal app, no VPN. Just open MyLLM. β ON-DEVICE OFFLINE MODE β No server? No problem. Download small language models (135M to 1.1B parameters) directly to your iPhone and run inference entirely on-device using Metal GPU acceleration. No internet or server connection required. Perfect for travel, commuting, or anywhere without WiFi. Available offline models: β’ SmolLM 135M β Ultra fast, works on any iPhone β’ Qwen 2.5 0.5B β Best balance for on-device use β’ TinyLlama 1.1B β Highest quality, iPhone 12+ The app automatically detects when your server becomes available and shows a comparison guide β so you can switch back to server mode for the full experience with one tap. β 60+ AGENTIC TOOLS β MyLLM isn't just a chatbot β it's an autonomous agent that can ACT on your behalf: β’ Web: Search the internet, fetch pages, get live RSS news headlines from 16 sources across 8 categories β’ Files: Read, write, edit, search, and organize files across your iOS storage β documents, downloads, cache, inbox β’ Contacts & Calendar: Search contacts, view events, create meetings, set reminders β hands-free β’ Photos: Browse albums, search by date, save images to your library β’ Device: Clipboard, notifications, device info, storage stats, haptic feedback β’ Shortcuts: Trigger iOS Shortcuts, open apps, get directions, share content β’ Translation: Instant text translation between 12 languages via your local LLM