Language Technology / Mobile Application · In development

Dilinga

Take the language barrier out of the conversation.

Dilinga translates the conversation itself. A message leaves in your language and arrives in theirs; a voice or video call is translated while it is being spoken; the camera reads the text in front of you and answers questions about it. And the translation is spoken in your own voice — the person on the other side hears you, not a stranger.

Messages arrive in the other language
Voice and video calls translated live
Text and speech translation
Camera translates what it reads
Ask the AI about the photo
Translation speaks in your own voice

A conversation between English and Italian: the sent message and the reply, both read in English, with the language pair in the header — You write EN, it arrives in IT

Dilinga's conversation list: four people, each with the language they write in, and every last message shown in the reader's own language — Every conversation in your language

The contacts list with the language of each person, and a voice and a video call button on every row — Call anyone in the list

The translation screen: an English question about breakfast and its Italian translation, each with a button to have it read aloud — Text and speech, either direction

The subway entrance sign at 34 Street Penn Station in New York, read through the camera: the app has laid the Japanese translation over the English text, each line in its own place — The sign, read and translated

The Ask screen with the same New York photograph loaded and the question box open: what do you want to know about this photo — Ask about the photo

The people and the messages are fictional and the New York photograph the camera reads is a stock image; the screens themselves are the app running, and the translations in them are its own.

How it works

Translation apps sit beside the conversation: you type somewhere else, copy the result and paste it back. Dilinga is the conversation — the message, the call, the photograph and the question about it all happen inside the app, so nothing has to be carried from one place to another.

Speech is the part people give up on first: a translated call in a stranger's synthetic voice stops sounding like you. Dilinga models the speaker's own voice, so the sentence that reaches the other side is in a language you do not speak but in the voice they know.

01
Write, they read it in their own language
The message is translated on its way out. Both sides write the way they always do, and each one reads their own language.
02
Call and be understood
Voice and video calls are translated as they are spoken — the sentence arrives while the conversation is still going.
03
Point the camera at it
A menu, a label, a form: the camera reads the text and translates it. The Ask screen answers questions about the same photograph.
04
In your own voice
Voice analysis carries the speaker's own voice into the translation, so every language is spoken by them rather than by a synthetic reader.

Ask about Dilinga.

It is still being built. Tell us which languages you need it in and what you would use it for — we will let you know the day it opens.

Dilinga

How it works

Write, they read it in their own language

Call and be understood

Point the camera at it

In your own voice

Ask about Dilinga.