Voice Dictation — Speak Instead of Type
Talk to Claude instead of typing. v0.8.0 adds on-device voice dictation to the composer: click the mic next to Send, speak, and your words land in the input box.
- Mixed-language recognition in a single turn — AAA Code transcribes Chinese, Japanese, English, Korean, and Cantonese freely mixed within one sentence. Say "こんにちは。请帮我 check 一下本地的 branch。감사합니다." — four languages in one breath — and every word lands correctly.
- Goes beyond Claude Code's
/voice— Claude Code's built-in dictation transcribes one language at a time (set via itslanguageoption; left empty, it defaults to English), so it can't follow a sentence that switches languages mid-stream. A Japanese speaker dropping English terms into Japanese, or anyone mixing Chinese and English, gets garbled output. AAA Code handles the switching natively. - Local and private — transcription runs entirely on your machine. No cloud service, no API key, no audio ever leaves your device.
- Hands-free flow — pause and the sentence drops into the composer; keep talking for the next one. Edit before you send, just like typed text.
Recognized languages: Mandarin Chinese, Cantonese, English, Japanese, and Korean — in any combination, within the same sentence.
First use downloads a one-time speech model (~230MB) on demand — nothing is bundled into the installer, so the download stays as small as ever.
Pricing: Basics Free Forever
We've simplified the model. The core experience — chat, unlimited sessions and workspaces, mobile remote access — is now free forever, no trial clock.
- Every new install still starts with a 7-day trial of all premium features.
- After the trial, premium features (Voice Dictation, Plugin System, Scheduled Tasks, Conversation Rename) unlock with a one-time purchase — VIP $10 or VVIP $15, with lifetime updates.
- The app no longer locks you out when the trial ends — you simply drop back to the free basics, with premium one click away whenever you want it.
