The archive · Consumer Apps · Product decision · 2016–2026
Timekettle bets dedicated earbuds, not apps, fix live translation
Shenzhen's Timekettle sells earbuds that interpret both speakers at once; a failed mic-array product nearly sank it, then a rebuilt line reached ¥200M revenue.
Timekettle
What the business is
Timekettle designs and sells consumer hardware for live translation — earbuds, handheld translators and conference terminals — and frames the product around the conversation, not the gadget.
How it started
Tian Li, a former Huawei gold-medal salesman, founded Timekettle in Shenzhen in 2016 after translation apps failed his parents on a trip to Europe: the apps needed button-pushing and turn-taking, so conversation kept stalling. The WT2 earbuds raised about $300,000 on overseas crowdfunding in 2017, and by 2019 they topped Amazon US's translation-earbud category.
What happened
Confident from early crowdfunding, the team mass-built the ZERO around a four-microphone array with mics spaced about one centimeter apart. In real far-field, multi-speaker rooms the beamforming and deep-noise algorithms broke down: complaints and returns piled up, tens of thousands of units sat in inventory, and cash flow came under pressure. Analyzing over 2,000 complaints, Timekettle rebuilt around clean pickup and two-way simultaneous interpretation — vector noise reduction on the W3 in 2021, bone-voiceprint pickup on the W4 — and launched Babel OS in January 2025, covering 52 languages and 106 accents with average latency under four seconds. Revenue passed ¥200 million in 2024 and kept growing in 2025, with more than a million cumulative users in 170+ countries.
How it ended up
As of September 2026 Timekettle is profitable and claims six straight years as the industry's number one; at IFA it launched the $379 W4 Plus, which translates YouTube and Netflix audio while the wearer uses one bud. Founder Tian Li's answer to Apple and Google: 'The logic of tech giants is to add translation functions to their existing ecosystems... our logic is to reconstruct the entire cross-language communication system.'
Background
Timekettle is a Shenzhen hardware company, founded in 2016 by former Huawei salesman Tian Li, that sells earbuds and devices for live translation between people speaking different languages. Its founding bet was that translation apps are the wrong interface for conversation: they force turn-taking and button-pushing, so Timekettle made earphones that interpret both speakers at once.
Early success — a $300K 2017 crowdfunding and the top Amazon US translation-earbud spot by 2019 — led the team to mass-produce the ZERO around a four-microphone array. The four mics sat about a centimeter apart, so beamforming and deep-noise processing collapsed in real far-field scenes: returns and bad reviews piled up, tens of thousands of units sat in inventory, and cash flow came under pressure. After analyzing 2,000+ complaints, it rebuilt around clean pickup and two-way simultaneous interpretation: W3 (2021) with vector noise reduction, W4 with bone-voiceprint, Babel OS in January 2025.
The rebuilt line made the company profitable: 2024 revenue above ¥200 million with growth continuing in 2025, more than one million cumulative users across 170+ countries, about 70% of revenue overseas, and six consecutive years as the industry's number one. In September 2026 the $379 W4 Plus added one-bud video translation of YouTube and Netflix, and Tian Li argued the gap with tech giants is strategic: they add translation to existing ecosystems, while Timekettle rebuilds the cross-language communication system.
What has to be true
- Translation apps interrupt conversation: each speaker must operate the phone and wait, breaking the natural two-way exchange the product promises.
- ZERO proved lab specs are not real rooms — a centimeter-spaced mic array failed in far-field, multi-speaker use, so the company now designs around far-field pickup.
- Giants treat translation as a feature of an existing ecosystem; Timekettle makes it the product, giving it a reason to optimize latency, accents and conversation flow.
- Dedicated hardware with its own protocol lets both speakers talk at once, which a translation app bolted onto a phone cannot reproduce.
What can be applied
Mic arrays that work in demos fail in real rooms, so Timekettle rebuilt around far-field pickup and two-way interpretation, betting dedicated hardware beats translation bolted onto an ecosystem.
Aftermath
As of September 2026 Timekettle is profitable, claiming six straight years as the industry leader with 2024 revenue above ¥200 million and over one million cumulative users. Its W4 Plus sells for $379 and translates streamed video through a single bud. The standing risk is the one Tian Li is betting against: Apple, Google and Samsung can ship translation as a free feature inside earbuds people already own, and the dedicated market is small — QYResearch sized simultaneous-interpretation earbuds at about $252 million in 2025 — so Timekettle must keep the experience better than a bundled feature.
Sources
- 38岁前华为金牌销售创业做翻译耳机，卖遍全球170多个国家，年入超2亿元
- Exclusive Interview with Timekettle's CEO: Unfazed by Apple's Entry, AI Translation Earphone Market Not Near Endgame
- Timekettle's new earbuds are better at translating videos
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card