Offline LLM vs Local LLM

Offline LLM runs an AI language model directly on your iPhone. Not a front end for someone else's chatbot — the model is downloaded to your device and every answer is generated there. That means it works in Airplane Mode, in a tunnel, on a flight, and in any building with bad reception. It also means nothing you type is sent anywhere, because there is nowhere for it to be sent. No account, no sign-up, no analytics, no server. WHAT YOU CAN DO Chat — Summarise, rewrite, explain, draft, brainstorm. Replies stream in as they are written. Your history is searchable and you can pin the conversations you keep coming back to. Ask your documents — Add PDFs or paste in notes. They are chunked and indexed on the device, and when you ask a question the answer cites the passage and page it came from, so you can check it. Ask a photo — Vision models describe images, read text out of them and pull out details. The photo never leaves your phone. Memory — Facts it remembers about you, shown in a plain list you can edit, with a switch on every single entry. Plus custom instructions for how it should write. Model lab — Live memory and thermal readout, speed benchmarks measured on your own device, and sampling controls. Open model library — Browse MLX-ready models on Hugging Face and load your own. PICK A MODEL THAT FITS A curated catalogue of open models — Qwen, Llama, Gemma, Mistral, Phi, SmolLM, Granite and more — from 76 MB up to 4.4 GB. Download only what you want and delete any of them at any time. Every model is rated against your specific iPhone before you download it. Not by looking up how much RAM the device has, but by reading how much memory iOS will actually let this app hold before it shuts it down. You find out whether a model runs before you spend the bandwidth, not after. SIMPLE OR ADVANCED Simple mode shows chat, documents and photos, and picks sensible settings for you. Advanced mode adds the model lab, speed measurements, sampling controls and the full Hugging Face library. Switch any time — nothing is removed, only hidden. HONEST ABOUT WHAT IT IS A model that fits on a phone is not a frontier cloud model. Offline LLM is genuinely good at summarising, rewriting, pulling details out of your own documents and everyday questions. It is weaker at hard reasoning and long chains of logic, and it knows nothing about recent events. Larger models on the device are slower but sharper — the app tells you the trade-off for each one instead of hiding it. REQUIREMENTS iPhone 11 or newer, iOS 17 or later. Inference uses the device's GPU, so older hardware is not supported. PRIVACY Your conversations, documents, photos and notes stay in the app's own storage on your device and are never uploaded. The app only reaches the network to download a model you asked for, and to process a purchase through Apple. Deleting the app removes everything it stored, models included. TRY IT, THEN DECIDE Everything is unlocked for three days. No payment details, nothing to cancel, nothing happens automatically. After that a small model stays free forever, and every conversation, document and note you already made stays exactly where it is — nothing you wrote gets locked away. Offline LLM Pro unlocks the full catalogue, document indexing, vision, memory and the model lab: Monthly — billed every month Yearly — billed every 12 months Lifetime — a single payment, no subscription Prices are shown in the app in your local currency before you buy. Subscriptions renew automatically unless cancelled at least 24 hours before the end of the current period. Manage or cancel in Settings > Apple Account > Subscriptions. Payment is charged to your Apple Account at confirmation of purchase. Terms of Use: https://www.apple.com/legal/internet-services/itunes/dev/stdeula/ Privacy Policy: https://patrickgerard.de/offlinellm/privacy/ Test everything free. Then: Monthly — $4.99 per month Yearly — $24.99 per year Lifetime — $49.99, one payment

Local LLM runs language models on your iPhone. No account is required. Chat, ask about photos, work with documents and use live voice. With Canvas, build websites, interactive pages and games from a conversation, then preview and edit them in the app. Features depend on the selected model. Download models from the catalog, or use Apple’s on-device model on supported devices when available. Core AI models require iOS 27 or later. Other supported models remain available on older systems. Inference and conversation history stay on your device. Model downloads, the catalog and model artwork use the network. Optional Web Search sends a query derived from your message to Tavily and may read result websites. Weather requests use Apple geocoding and WeatherKit for the requested city, without requesting device location. Voice input prefers on-device recognition. If it is unavailable or fails, Apple’s online speech service may process audio when network speech recognition is enabled. You can disable this in Privacy settings. Requires iOS 17 or later. Model availability depends on your device, OS, memory and storage. Downloads require internet; supported offline chat works after setup. Models can make mistakes. Do not rely on replies for medical, legal or financial decisions.