Why written text is the bigger problem, not speech

Conversations are occasional. Signs are constant. Between arriving at an airport and getting to a hotel, a traveler passes dozens of pieces of written information — directions, warnings, restaurant names, transit boards — almost all of it untranslated and, in much of Southeast Asia, in a script that can't be typed on a standard phone keyboard even if you wanted to. Camera translation exists specifically for this: point, and the text on the screen appears translated in place, without ever needing to type a character of a language you can't input.

How it actually works

The camera captures the text in the frame, on-device text recognition identifies the characters (whether Latin, Thai script, or Chinese characters), and translation overlays the result directly over the original — so you see a menu that reads in your language while still looking like the menu, rather than a separate block of translated text disconnected from the layout.

This matters more than it sounds: a translated menu that keeps its original structure (dish name, price, description in roughly the same spot) is far easier to actually use while ordering than a flat list of translated sentences in a different order.

Hand holding a smartphone using camera translation to read a Thai street food menu board, next to the real board
The translated text overlays the original, in the same layout.

Where it's most useful

What to keep in mind

Camera translation works best with clear, well-lit, reasonably flat text — a printed menu or a shop sign, rather than a curved surface or heavily stylized calligraphic font. Handwritten specials boards vary in legibility just as they would for a human reader, so results improve with steady framing and good light. For anything ambiguous or important (a medical instruction, a legal notice), treat the translation as a strong guide rather than a certified one.

Camera and voice work together

Reading a menu answers "what is this." Ordering it, asking about spice level, or requesting a substitution still needs a spoken exchange. The two modes are complementary — camera gets you through the written half of a foreign environment, voice gets you through the spoken half — and switching between them on the same trip, sometimes the same minute, is the normal way travelers actually use both.