Help & support.
Direct answers about capture, on-device analysis, and what leaves your Mac.
Capture, ask, read.
hotkey · preset · answerNo setup required. Try the bundled sample image first — it needs no permissions at all.
Capture.
Press Option-X to drag-select any part of any display, use the camera, open an image file, or start with the bundled sample.
Ask.
Pick the preset that matches what you are looking at and type your question.
Read.
The answer is computed on your Mac, labelled with its source, and copied for pasting if you want it to be.
Vision, stated honestly.
on-device · providers · permissionsThe app tells you what it detected and how confident it is. It does not invent answers it cannot support.
Press Option-X anywhere (or choose “Explain Screen Region” from the menu bar icon), drag a box over the part of the screen you care about, and release. The capture opens in the floating panel — pick a preset, type your question, and press Analyze. The first time, macOS will ask you to allow Screen Recording in System Settings; that permission is only used when you invoke a capture.
All image reading is on-device: Apple's Vision framework performs text recognition, image classification, and barcode/QR decoding. On macOS 26 or later with Apple Intelligence enabled, Apple's on-device language model writes the answer from those detections — still on your Mac, and the model never receives the image itself, only the detected text and labels. Nothing requires the internet unless you explicitly connect a cloud provider.
That is the honest fallback. Without Apple Intelligence (earlier macOS, unsupported hardware, or the feature switched off), Ohms Explain reports exactly what it detected — recognized text with its measured confidence, likely subjects, decoded codes — rather than inventing a summary it cannot support. Connect a provider of your own, or use a Mac with Apple Intelligence, for written answers.
Choose “Explain from Camera” from the menu bar. You can pick a built-in camera, an external one, or your iPhone via Continuity Camera. macOS asks for Camera permission the first time. The shutter hands the photo straight to the analysis panel; photos are analyzed in memory and never written to disk.
No. There is no account, and the app is fully functional offline. Providers are optional: if you want a frontier model to look at an image, connect Anthropic, OpenAI, Google Gemini, OpenRouter, or GitHub Models with your own API key, or point the app at an Ollama server on your own machine.
Only when you select that provider and press Analyze: the captured image and your question go straight from your Mac to that provider's API under your own key. Silo Systems runs no server and receives no copy. Each cloud provider row is labelled SENDS IMAGE so this is never ambiguous. Read your provider's own retention policy.
In the macOS Keychain, in app-scoped items. Never in preferences, files, or logs, and never transmitted to Silo Systems. Disconnecting a provider removes its key.
History stores your question, the preset, the answer text, and its source — never the image. It lives in a single local file in the app's Application Support container. You can search it, delete any entry, or clear it entirely from the History window.
No. It reads what is printed on packaging — ingredients, allergens, weights, dates, codes — and every answer it produces carries an informational-only disclaimer. It is not a substitute for professional medical, dietary, or safety advice.
A one-time Mac App Store purchase at $29.99. No subscription, no credits, no account. Any provider you connect bills you directly under your own key.