OCR & detection
How Oniro turns a photo or camera frame into cards, positions, and reading context.
What detection reads
Oniro looks for card boundaries, spread order, card titles, and orientation. The final reading uses both the detected card identity and the meaning of its position.
What affects accuracy
Glare, blur, low contrast, cropped cards, heavy shadows, and overlapping layouts make detection less reliable. Decorative decks can also be harder to read when the title area is small or stylized.
Why position matters
The same card can mean different things in different slots. "Obstacle", "advice", and "near future" are not interchangeable, so check the layout before accepting the reading.
Manual edits are expected
Detection is designed to save time, not remove judgment. Edit card names, reversals, or positions whenever the preview does not match your table.
Privacy note
Only upload the spread image you want interpreted. Avoid including faces, personal documents, or anything in the background that is not part of the reading.