SiloLM

SiloLM runs open language models directly on your iPhone, keeping everything you say private and working even in airplane mode. Core capabilities: • Fully on-device processing with no account or cloud required • Image understanding for photos, charts, and screenshots • Real-time web search with cited sources • Summaries and answers from any URL you paste • Drafting, rewriting, and polishing texts inside iMessage • Adjustable reasoning depth for quick replies or step-by-step analysis SiloLM works by running models accelerated by Apple silicon and the Neural Engine on your device. There are no servers, API keys, or shared queues. Your photos, prompts, and messages never leave the phone, and the app collects no data. You can choose from a curated library of open models, from lightweight options under 1 GB to multimodal models that read and reason about images, then switch between them whenever you like. The search tool lets the local model look up news, prices, and current events when you enable it, and it cites its sources so you can verify answers. Paste an article or link and SiloLM will digest it on demand. The reasoning control lets you trade speed for depth, getting instant replies when you are in a hurry or detailed, step-by-step thinking for harder problems. SiloLM is built for privacy-conscious users, professionals handling sensitive material, travelers who need offline access, and anyone who wants cloud-level chatbot features without sending their data elsewhere. Use it to brainstorm, summarize documents, review photos, write better messages, or research topics on the go.