01
Put an object in front of it
Talki sits on the table and looks ahead. Place something in its field of view and it recognises it and says the word.
Tangible interaction · [DA COMPLETARE: contesto, es. MAS SUPSI] · 2019
A device that lets children learn words in another language by pointing it at things around them. Point it at a mug, a pen, an apple — Talki recognises the object and says the word out loud.

Children learn a language faster when the word is attached to something real. A picture of a mug in a textbook is one thing; the mug on the table in front of you is another.
So the question was: what if learning a word started from the object rather than from the page?
Talki is a physical interface with a camera and an internal speaker. You point it at an object, it recognises what it is, and it pronounces the word in the language you're learning. No screen, no typing, no menus — you aim it at something and listen.
It's polyglot: the same object can be named in several different languages.
01
Talki sits on the table and looks ahead. Place something in its field of view and it recognises it and says the word.
02
Small enough to hold in two hands, so a child can carry it around the room and point it at whatever they find.
03
The answer comes as sound, not as text. The child hears how the word is pronounced, which is the part a written translation can't give.

The size was a design constraint, not an afterthought. A device a child can hold in two hands and carry across a classroom behaves differently from one that stays on a desk: it turns looking for objects into part of the lesson.


The use we designed for is a teacher introducing a topic. Instead of listing the words, the teacher asks the class to go and find objects related to it, and Talki names each one as the children track it down.
The learning happens through searching and pointing rather than through repeating, and the words stay attached to the objects they came from.

[DA COMPLETARE: le decisioni reali, con il trade-off dichiarato per ciascuna. Le tre che mi aspetterei, ma che devi confermare o correggere tu:]
[DA COMPLETARE: perché un oggetto dedicato invece di un'app su telefono, che avrebbe già camera e altoparlante. Trade-off?]
[DA COMPLETARE: perché solo audio, senza schermo. Cosa ci avete guadagnato e cosa ci avete perso?]
[DA COMPLETARE: la tecnologia di riconoscimento. Modello pre-esistente, servizio cloud, training vostro? E funzionava davvero o era simulato?]
[DA COMPLETARE: l'esito onesto. È stato testato con bambini veri? È rimasto un prototipo? Cosa avreste fatto dopo, e cosa non ha funzionato?]

