Speak once.
Paste polished text.
Hold a hotkey, speak Mandarin or English, and Kiao Voice cleans it up with your own terminology — then pastes it into whichever app you're in. No tabs, no dialogs, no friction.
Built in Taiwan · Apple Silicon · Free during beta
From thought to text in three steps
Hold the hotkey
Press your configured shortcut from anywhere on macOS. A subtle overlay confirms the mic is hot — no window switching needed.
Speak naturally
Mix Mandarin and English however you think. Kiao captures your audio with on-device privacy defaults and high-accuracy STT.
Paste polished text
AI enhancement fixes grammar, applies your custom dictionaries, and pastes the result directly where your cursor was.
A Taiwan product built for the way you actually work
Most voice-input tools come from somewhere else. Kiao is built here — with the trust model, the language tuning, and the business plumbing that Taiwanese teams care about.
Your data stays in trusted hands
Kiao Voice is a Taiwan-based product. Not a Chinese company, not re-branded from one. Your recordings and transcripts flow through infrastructure you can verify — not across jurisdictions you cannot.
Tuned for how Taiwan actually speaks
Vocabulary, phrasing, and common typos are optimized for Traditional Chinese and Taiwanese usage — 捷運, 拜託, 有沒有, code-switching with English. Less post-editing, more flow.
Taiwan business invoices, first-class
Need a 統編 invoice for your company books? We can issue one. Most overseas voice-input tools can't — so their paper trail stops at a credit card receipt.
Made for teams that live in two languages
Bilingual by design
Mix Mandarin and English mid-sentence. Kiao keeps the rhythm of how you actually think and write.
Custom dictionaries
Feed your team glossaries, product names, and domain terms. Every transcription respects them automatically.
Hotkey-first flow
Global shortcut from any app. Your cursor never leaves its place — the paste lands exactly where you were typing.
Context-aware cleanup
AI enhancement fixes filler words, casing, and grammar while preserving tone, code blocks, and technical terms.
Private by default
Audio is transcribed through your configured provider and discarded after processing. No unnecessary storage.
Built for speed
Tight feedback loop from recording to paste — no dialogs, no multi-step wizards, no context switching.
Start speaking your best writing.
Free during beta. Works on any Apple Silicon Mac.
Latest · v0.2.4 · Apr 12, 2026