Translation Earbuds App: Each Person Hears Their Language (Without Buying New Earbuds)

Timekettle sells translation earbuds for $249 to $699. The pitch is compelling: each person wears one earbud from the same pair, and each hears the other person’s speech translated into their own language, privately, in their ear. No phone screen between you. No loudspeaker broadcasting your conversation to everyone nearby. Just a natural conversation across a language barrier, handled quietly by the hardware in your ears.

It’s a genuinely good product. It’s also $249 at the low end — and the feature it delivers is available in software, on the earbuds you already own, starting free.

That’s what Puente’s Earbud Share Mode does.

What “Translation Earbuds” Actually Means

The use case people are searching for when they type “translation earbuds” or “earbud translation app” is specific: a two-person conversation where both parties stay in their own language throughout, each hearing the other’s words translated privately in their ear, with no shared screen and no speaker output audible to anyone else.

This is meaningfully different from holding a phone between two people and having a translation app read out loud — which works but feels like a parlor trick and eliminates any privacy. It’s also different from traditional interpreter apps where one person speaks into the phone, taps a button, and hands it to the other. That works for simple exchanges but breaks down in any conversation that has a natural rhythm.

The earbud version — each person in their own language, each hearing the translation privately — is close to how a real human interpreter works. One person speaks, the other hears the translation immediately in their ear, responds in their language, and the first person hears that response translated back. The conversation flows. That’s the experience people are looking for.

Why Dedicated Hardware Was the Only Option (Until Recently)

Timekettle and similar hardware products emerged because delivering this experience reliably required solving a latency problem and a channel separation problem simultaneously. Each earbud needed to receive a different audio stream — the translation of what the other party said — with low enough latency to feel conversational rather than robotic.

In the early days of translation apps, doing that in software on a single phone wasn’t feasible. Audio APIs didn’t expose the level of channel control needed, processing overhead introduced too much latency, and Bluetooth audio routing wasn’t reliable enough for real-time channel-split output.

Dedicated hardware sidesteps the software problem entirely: each earbud connects to a separate device, processes audio independently, and delivers output to its wearer. The hardware cost is the price of reliable execution.

What changed is the software stack on modern iOS. Current audio APIs give developers direct control over stereo channel output with sub-second processing. Puente routes left and right channel audio to the correct earbud in real-time, entirely in software, on a single iPhone. The earbuds don’t need any firmware update, any proprietary connection, or any hardware-side processing. They just need to be a stereo pair — which every pair of earbuds sold in the last decade is.

The hardware dependency is gone. The software does it now.

How Puente’s Earbud Share Mode Works

The setup takes about thirty seconds:

1. Open Puente and select Earbud Share from the mode selector
2. Assign languages — Party A goes to the left channel, Party B to the right
3. Each person takes one earbud and puts it in

From that point, the session is bidirectional and continuous. Puente picks up both speakers through the iPhone microphone — no push-to-talk, no tap to switch. Party A speaks; Party B hears the translation in their earbud. Party B responds; Party A hears that translation in their earbud. The phone sits between them or in a pocket.

The translation engine is DeepL Voice, which scores 96.4/100 on translation quality benchmarks — meaningfully better than Google’s 87–89. For conversations where accuracy matters (medical, legal, business), the quality difference is audible.

For 8 languages — English, Spanish, French, German, Portuguese, Italian, Japanese, and Mandarin — the entire session can run offline using on-device Whisper AI. No internet connection needed, no audio sent to any server. For the remaining 101 languages, audio is processed by DeepL Voice and immediately discarded. No account required either way.

The Cost Comparison

If you own any stereo earbuds — AirPods, Galaxy Buds, Sony WF, Jabra, Bose, or a $25 generic pair — you can start translating for nothing. No $179 device. No $699 device. And if you outgrow the free tier, a $2.99 Day Pass (120 minutes) or a minute pack (from $4.99) keeps you going with no subscription, while the Plus plan ($9.99/mo) covers regular use.

If you don’t own earbuds and need to buy some, AirPods 4 ($129) plus Puente gets you the same experience as a $699 Timekettle W4 Pro for the price of the earbuds alone — with 130+ languages versus Timekettle’s 40, offline support in 8 languages, and the Empathy Engine for emotional tone transfer that no hardware translator offers.

Use Cases

Doctor and patient. An English-speaking physician and a Spanish-speaking patient split a pair of AirPods. The physician speaks, the patient hears Spanish in their ear. The patient responds in Spanish, the physician hears English. The consultation happens at a natural pace with no interpreter on hold and no phone screen as a barrier. For the 8 offline languages, nothing leaves the device — HIPAA-aligned for even the most sensitive clinical environments.

Attorney and client. A lawyer and a Mandarin-speaking client in a conference room. Earbud Share Mode means nothing is broadcast to anyone nearby. The conversation stays between the two of them. Voice Identity (Acoustic Compass) attributes each turn to the correct speaker in the Live Captions overlay, producing a labeled record of who said what.

Bilingual couple. One partner speaks English, the other speaks Portuguese. They’ve been switching languages awkwardly for years — whoever speaks the other person’s language better ends up carrying the weight of most conversations. With Earbud Share Mode, each person speaks their own language throughout. No translation burden on either party.

Traveler. Someone visiting Japan without Japanese skills approaches a local, hands over one earbud, and conducts a real conversation — asking for a restaurant recommendation, getting actual directions, discussing something beyond what pointing and nodding can communicate. The phone stays in one hand. The conversation flows.

No Loudspeaker: Privacy by Design

One detail worth calling out directly: in Earbud Share Mode, translated audio goes only to the earbuds. There is no loudspeaker fallback. Nothing plays through the phone speaker.

In a medical exam room, a legal office, or anywhere a private conversation is happening, this matters. A translation app that reads the conversation out loud broadcasts confidential information to anyone nearby. Earbud Share Mode routes everything to earbuds. No one else hears anything.

Combined with no audio storage, no account requirement, and offline processing for the 8 most-used languages, Puente is the most privacy-forward translation option available — and it happens to also be the cheapest way to get earbud translation.

The No-Hardware Case

Timekettle is a good product. If you want standalone hardware that doesn’t depend on a phone, or you want physical buttons instead of a touchscreen, or you just prefer a dedicated device, it’s a reasonable buy at the M3 price point.

But the feature itself — each person hearing the other’s translation privately in their earbud — is not hardware. It’s a software capability. Puente has it, it works with the earbuds you own, and you can start using it for free — with a $2.99 Day Pass or the Plus plan ($9.99/mo) there when you need more.

The translation earbuds experience is no longer something you buy. It’s something you download.