Translation apps help you follow what is said, then leave you on your own. EchoBridge helps with both halves: you hear them in your language, in your own headphones, as they speak. Then three replies appear in theirs. Tap one and your iPhone says it.
In TestFlight, App Store to follow.
Buenas, ¿ya saben qué van a pedir?
Hi, do you know what you'd like to order?
El especial de hoy es ajiaco.
Today's special is ajiaco.
viene con arroz y aguacate…
A café in Medellín. Grey italic is still being heard.
The other half is answering, and that is where most translators stop. EchoBridge is built for both.
EchoBridge doesn't wait for the end of a sentence. It translates in short phrases as they settle and speaks each one into your headphones: AirPods, Bose, or any other Bluetooth pair. The original and the translation stay on screen, so you can read along when it's noisy.
When they finish, three short replies appear in their language, each with its meaning in yours: one to say yes, one to ask something, one to decline politely. Tap one and your iPhone speaks it from its own speaker, so they hear it and your earbuds stay in.
A suggestion is a starting point, not a script. Nothing is said until you choose it.
Buenas, ¿ya saben qué van a pedir? El especial de hoy es ajiaco.
Hi, do you know what you'd like to order? Today's special is ajiaco.
Replies are written for the conversation so far, in polite, everyday spoken language, and kept to twelve words or fewer so they are easy to say or to tap.
Three replies, each a different intent, with the meaning in your language underneath. They follow the last few exchanges, not just the last sentence. Tap one and your iPhone says it.
Touch and hold a suggestion to change a word first: a dish, a time, a name. The meaning stays beside it, so you know what you're about to say.
Rather say it yourself? Hold the button and speak in your language. Let go, and EchoBridge translates it and says it out loud in theirs.
For the quiet moments. Type in your language, preview the translation, adjust it if you like, then tap Speak.
Your replies come from the iPhone's own speaker, toward the person you're talking to. Your earbuds stay connected, and EchoBridge goes back to listening as soon as it's done.
EchoBridge never speaks a suggestion on its own. If you'd rather just listen, turn suggestions off in Settings and nothing appears.
Four steps, and the first one you do once.
Pick the language they speak and the one you speak. Spanish and English to start; the Languages screen lists every pair your iPhone can handle on device.
Headphones in, or not. iOS shows its microphone indicator whenever EchoBridge is listening, and Stop ends it.
Short phrases arrive in your language as they're spoken, in your ears and on screen, with the original underneath.
Tap a suggested reply, hold to talk, or type. Your iPhone says it in their language, then goes back to listening.
Switch at the top of the screen, any time.
Ordering, asking the way, checking in, chatting with a neighbour. Translations in your ears, suggested replies, Hold to Talk and typed replies. Swap the languages with one tap when the roles change.
A tour, a meeting, an announcement, the family at the next table. No suggestions and nothing spoken back. Choose Text only to read silently, or In earbuds to hear it. A thin line marks a pause that may be a new speaker.
Built on the iOS 27.1 SDK, so it fills the Duo's displays edge to edge and lays itself out for however the phone is held: closed, open, set down like a laptop or stood up like a tent. Every orientation.
On the outer display EchoBridge is the phone layout: transcript above, suggested replies and the Start and Hold to Talk buttons at your thumb, with History, Settings and Type a reply in a bar along the edge.
Open it flat and the conversation gets a side of its own: what they said on one half, your replies, Start and Hold to Talk on the other, with nothing sitting on the fold.
Set it half open like a laptop and the transcript takes the upright half while your replies and buttons rest on the flat half nearest you. Stand it up like a tent and the conversation faces you on one half while it keeps listening.
EchoBridge uses Apple's on-device speech recognition, translation and language model. There is no EchoBridge server to send anything to.
| What | Where It Happens |
|---|---|
| Hearing what's said | On your iPhone — Apple's on-device speech recognition |
| Translating it | On your iPhone — Apple Translation, with language packs that download once |
| Suggesting replies | On your iPhone — Apple's on-device model, on iPhones with Apple Intelligence |
| Speaking for you | On your iPhone — the voices installed on it |
| The audio itself | Never recorded — listened to as it arrives, then gone |
| Transcripts | Not kept — unless you turn on Keep transcripts; then only on this iPhone, and deletable at any time |
An iPhone, and ideally a pair of headphones you already own.
| Item | Requirement |
|---|---|
| iPhone | iOS 26 or later. Ready for iPhone Duo, closed or open |
| Headphones | Optional. Any Bluetooth headphones: AirPods, Bose, Sony and the rest. Without them, read along on screen |
| Suggested replies | An iPhone with Apple Intelligence turned on (iPhone 15 Pro or later), in a language Apple's model supports. Without it, everything else works and suggestions are simply absent |
| Languages | Any pair your iPhone supports on device for both speech and translation. Spanish ⇄ English is the default and the pair it's tuned for |
| Internet | Once — to download a language's speech model and translation pack. After that, no |
For a more natural voice, download an Enhanced or Premium voice in iOS Settings › Accessibility › Spoken Content › Voices, then pick it in EchoBridge's Settings › Voices.
It is not a simultaneous interpreter. It translates phrase by phrase, so you hear each phrase a moment after it's said. When someone talks faster than the voice can keep up, older audio is skipped and the text stays on screen.
Use it to get through a conversation, not to sign something. For medical, legal or emergency situations, find a human interpreter.
It can't separate two people talking at once, and its suggestions are only as good as what it heard. It runs on iPhone, including iPhone Duo, only for now: there is no iPad, Mac or Apple Watch app in this version.