Repository navigation
Tidytalk 0.2.0
- 更多识别模型:在"设置 › 识别模型"里可以换成 Qwen3-ASR(最准,能直接认出词库里的人名和术语)、Fun-ASR-Nano、Whisper Turbo,或 macOS 自带的识别。默认仍是 SenseVoice,所有模型都在本机运行。
- Apple 智能:支持的 Mac 上可以在"配置 AI"里选 Apple 智能整理和翻译,不需要 API Key,文字不出电脑。
- 转写文件和链接:菜单里选"转写文件或链接…",拖入音频、视频文件或粘贴 YouTube、B 站链接,得到带时间的文字,可导出 TXT、SRT 字幕或 Markdown。
- 自动学词:粘贴后手动改过的词,会提示你一键加入词库,确认后才加入。
- 实时字幕边听边翻译:不用等一句话说完,译文跟着原文出来。从某一句粘贴改为悬停后点"粘贴"按钮,点句子或拖动字幕条不会误粘贴。
- 界面整理:菜单栏菜单按用途分组,设置合并为 6 页。
安装:需要 macOS 26 或更高版本、Apple 芯片的 Mac。安装包没有经过苹果公证,第一次打开请到"系统设置 › 隐私与安全性"点"仍要打开",或在终端执行 xattr -dr com.apple.quarantine /Applications/Tidytalk.app。
- More speech models: in Settings › Speech Model, switch to Qwen3-ASR (most accurate, and it recognizes names and terms from your vocabulary), Fun-ASR-Nano, Whisper Turbo, or the speech recognition built into macOS. SenseVoice is still the default, and every model runs on your Mac.
- Apple Intelligence: on Macs that support it, choose Apple Intelligence in Set Up AI to tidy and translate with no API key, keeping your text on your Mac.
- Transcribe files and links: choose "Transcribe File or Link…" in the menu, drop in an audio or video file or paste a YouTube or Bilibili link, and get timestamped text to export as TXT, SRT subtitles, or Markdown.
- Words you correct: after pasting, words you change by hand are suggested for your vocabulary, and nothing is added until you confirm.
- Live captions translate as you listen, without waiting for the sentence to end. Pasting from a sentence now uses a Paste button that appears on hover, so clicking or dragging the caption bar never pastes by accident.
- A tidier interface: the menu bar menu is grouped by purpose, and Settings now has six tabs.
Install: requires macOS 26 or later on a Mac with Apple silicon. The app isn't notarized. The first time you open it, click "Open Anyway" in System Settings › Privacy & Security, or run xattr -dr com.apple.quarantine /Applications/Tidytalk.app.
SHA-256 (Tidytalk-0.2.0.dmg): d2340d26a8c8dd6f950e0403dbde665e1aa3be28364ca0995be20981de19289b