What Does This Mean?

A sentence you can't parse. An error message. A screenshot. You don't have to switch apps, and you don't have to describe the situation all over again. What Does This Mean? starts from what you deliberately share. Inside the iPhone share sheet it produces one explanation of that content, in a fixed output structure: • Topic • One-line answer • Plain-language breakdown • Example • Key points • Follow-up questions You can keep asking along the same passage. The previous answer stays in context, so you never restate the background. HOW IT WORKS • It receives the text you selected, or a single screenshot, from the system share sheet. • Text recognition and image cues for screenshots are produced on your device. When the evidence is too thin to support an explanation, it stops and says so instead of guessing an answer for you. • Before content leaves the device, the app matches a set of secret patterns locally — private keys, certificates, tokens, API keys, database connection strings, password assignments, email addresses, phone numbers, national ID numbers — and masks them in place. This masking applies to text. • Your understanding ledger stays on your own device. The app does not read the clipboard, does not ask for Photo Library access, and does not capture your screen or watch other apps in the background. • No ads, no third-party trackers, no behavioural analytics, no account, no sign-in. WHO PRODUCES THE EXPLANATION Apple's on-device model is the default and needs no setup — provided your iPhone supports Apple Intelligence, has it turned on, and has finished downloading the model. If your device does not qualify, the app tells you which condition is missing. You can instead use your own API key (currently OpenAI, Anthropic and xAI; the provider list may change between versions). Redacted text then goes to the provider you picked. The key stays in the local Keychain (ThisDeviceOnly): not in the ledger, not synced to iCloud. Letting that provider look at the image itself is a separate consent: the app asks for that share alone, and only after you agree does it send a metadata-stripped JPEG with its long edge capped at 1280 — never the original file. Declining costs you nothing; the app answers from the on-device cues instead. OUTPUT LANGUAGES Traditional Chinese, English and Japanese, following the system or set by you. WHAT TO KNOW FIRST Explanations are produced by a generative language model, and the same input may produce different output at different times. What this app performs is this: it produces one explanation of what you shared, in the language you choose, in the fixed output structure above. It does not promise that any particular explanation is accurate, complete or suitable. An explanation is an aid to understanding — not legal, medical, financial, immigration or tax advice. For anything that matters, check it against the source. Image understanding is beta. REQUIREMENTS iOS 26 or later. The zero-setup on-device path needs an iPhone that supports Apple Intelligence with it turned on; otherwise you supply your own third-party API key.