In a launch event at IFA 2025 in Berlin, Timekettle unveiled the W4 AI Interpreter Earbuds – the company’s latest attempt to make seamless cross-language communication a reality. The new W4 buds are built around two core innovations. First is the bone-voiceprint sensor, which captures voice vibrations directly from the user’s bones before speech leaves the mouth. That means the system can filter out ambient chaos – whether you’re in a subway, an airport, or a bustling trade show – and still deliver crystal-clear input.
Second is Babel OS 2.0, the company’s AI translation platform. Powered by advanced LLM models, Babel OS can interpret context, fix homophone errors, and provide translations in real time with what Timekettle claims is 98% accuracy and just 0.2 seconds of delay. Together with the bone conduction technology, the Timekettle is able to retain greater than 90% accuracy even in 100dB environments.
The earbuds support 42 languages and 95 accents. Modes like one-on-one conversation and listen-and-play make them suitable for personal interactions, meetings, and even multilingual family conversations. Battery life runs up to 10 hours with the charging case, and they double as high-quality music earbuds.
During an expert panel at the launch event, Timekettle CEO Leal Tian described the importance of these improvements. “Reducing latency is not just a technical milestone, it changes the way people interact. A near-instant response makes conversations flow naturally, almost as if both parties were speaking the same language”. He also teased an upcoming feature that will let the system clone the speaker’s tone of voice, preserving emotion and nuance across languages.
Prof. Dr. Thomas Kosch, a Human-Computer Interaction Specialist at Humboldt University who was on the panel, noted that while “human interpreters are still irreplaceable in sensitive cultural or diplomatic contexts…in the future, both will complement each other – AI for everyday communication, humans for depth and nuance.”
The panelists also looked ahead to where technology like this might go. Ideas ranged from mid-sentence translation that could handle overlapping conversations to confidence ratings that would help users know when to trust translations fully and when to double-check. For Timekettle, the long-term vision is what Tian calls “L4 or L5 stage translation” – a future where conversation flows across languages as naturally as if everyone were speaking the same one.
I had the opportunity to test the earbuds firsthand, translating between English and Chinese with a company rep, and even on the noisy show floor they were impressive for their accuracy. Conversations felt natural, and the clever magnetic case design – which snaps apart so you can easily hand one earbud to the person you’re speaking with – makes them especially practical. The earbuds themselves are lightweight and comfortable enough for extended wear.
And I expect that the translation quality will continue to improve. Timekettle’s CTO Ryan Shi noted that the W4s come with lifetime software upgrades and that, over time, they will learn from your interactions, adapt to your context, and become increasingly personal.
At launch, the Timekettle W4 are available in Midnight Blue and Sandy Gold, priced at $349 / €349 through Timekettle’s website and Amazon.
