hex

anomalycoMIT活跃维护

跨 macOS、Windows、Linux 的本地语音听写工具,用 Rust 编写,把说的话实时转成文字,全程离线运行,不上传云端。

Local voice dictation for mac, windows and linux

v3.3.0发布于 2026-11-02 16:43 UTC6,043star +361 / 7 天370fork同步于 2026-11-10 11:34 UTC

怎么上手

由 README 提炼,执行前请确认
brew install --cask anomalyco/tap/hex
硬件要求按 README 说明

README 未说明具体显存或 CPU 要求;仅列出支持平台:macOS(Apple silicon)、Windows x64/ARM64、Linux x86_64/ARM64。

支持 Apple Silicon
下载:源码 zip·GitHub Releases(安装包)本站不镜像安装包

项目说明

由 README 自动翻译提炼

本地语音听写工具,按住快捷键说话,松手即转写并粘贴到当前应用

跨 macOS、Windows、Linux 的本地语音听写工具,用 Rust 编写,把说的话实时转成文字,全程离线运行,不上传云端。

核心能力

  • 按住快捷键说话,松开后自动转写并粘贴
  • 转写在设备本地完成,音频不出机器
  • 首次运行自动下载语音模型,之后离线可用
  • 提供 macOS、Windows、Linux 多平台安装包
  • 实验性功能集中在侧边栏底部的 Labs

容易踩的坑

首次运行需联网下载语音模型,之后才能离线使用;项目不接受 PR,只能开 issue 反馈;Linux 需按发行版选择 .deb、.rpm 或可执行文件。

内容由模型根据 README 生成,可能不完整或不准确,以官方文档为准。看原文 README

最近发版

5 个版本v3.3.0
  1. v3.3.0Hex 3.3.0最新正式版2026-11-02 · 2026-11-02 16:43 UTC

    Hex 3.3.0

    new

    • the first run has a quiet skip on the demo and practice, and if the speech model download fails or stalls you can carry on and pick a model later
    • the speech model picker compares three models side by side, then shows every model
    • russian: gigaam v3, and you can add your own gigaam model files
    • the interface speaks traditional chinese (繁體中文)
    • shortcut mode is one choice in general, and hex only offers the modes your desktop can deliver
    • each mode can have its own vocabulary
    • hex explains the desktop's paste permission before the desktop asks for it
    • when you set up a newer hex, an older copy that's still running quits by itself
    • mac: organizations can lock features on or off with a configuration profile
    • linux: apps can start and stop dictations through hex's local api, like on mac
    • new in labs: skip long pauses before transcribing (faster on slow computers), stop a meeting recording 30 s after the call ends, and fix a wrong result right from output's try box

    fixes

    • mac: the shortcut works again after cmd+tab or cmd+q left a key looking held, after hex sat idle for a while, and when it's fn
    • windows: dictation starts on mics that got stuck on "enabling…", and stream deck and other injected shortcut presses no longer get cut to nothing
    • dictation no longer fails with "the current clipboard format cannot be preserved safely"
    • spoken lists and numbers stay as you said them when hex cleans up your text
    • clearing or deleting history sticks, even after a failed try
    • if your opencode 2 is too old for hex, hex tells you to update it instead of skipping the ai rewrite without a word
    • linux: hex picks up settings you edit outside the window, the status and indicator name the mic in use and warn when it goes away, and hex no longer crashes at startup on gnome + nvidia under xwayland
    • kde: hex shows up in the shortcuts settings, waits for you to approve the shortcut, and asks again if you didn't. the default is ctrl+alt+space
    • gnome: the dictation shortcut registers reliably
    • niri and other desktops whose shortcuts portal can't bind: hex still starts listening, and hex dictation works from your own keybinding
    • annotate: the pen works on hyprland and on fractional scaling, saved ink lines up, recording works on intel + nvidia laptops, and captures finish cleanly
    • custom models that failed to warm up on the gpu load again
    • the speech engine row no longer says the gpu stopped responding when it's fine
    • meetings: the ready and recording capsules have a close button
  2. v3.2.0Hex 3.2.02026-11-01 · 2026-11-01 22:43 UTC
    查看更新说明

    Hex 3.2.0

    new

    • a new first run on every os: hex sets itself up and downloads the speech model for your language, then you try it by answering a few practice messages with your shortcut
    • if hex finds speech models already on your computer (in the hugging face cache), the first run lists them next to the recommended one in one simple list, so you can skip the download
    • models: hebrew (ivrit.ai's whisper large v3 turbo) and nemotron 3.5 asr streaming 0.6b
    • the interface speaks brazilian portuguese (português (brasil))
    • the microphone is one switch beside the picker, and settings say what each microphone mode means for privacy
    • speech engine has an auto / gpu / cpu setting, and hex remembers a gpu that hung so auto stays on the cpu for it

    fixes

    • dictation no longer gets cut off after 2.5 seconds on mics that add a steady dc offset, like some built-in laptop mics (#841). 3.1.2 fixed this for quiet mics
    • escape no longer cancels a dictation, so a stray key can't throw away what you said
    • windows: dictations no longer pile up in clipboard history (win+v), and pasting works in remote desktop sessions
    • linux: kde wayland hotkeys work when the portal starts before plasma, the shortcut recorder takes home, end, page up/down and arrows, the hyprland shortcut docs are current, and annotate records video on intel + nvidia laptops
    • mac: the recording indicator lands in the right place on a second display, and gather calls get detected along with arc, meet and whatsapp
    • long menus and pickers show that they scroll
    • the homepage at https://anoma.ly/hex is faster
  3. v3.1.2Hex 3.1.22026-11-01 · 2026-11-01 16:56 UTC
    查看更新说明

    Hex 3.1.2

    Fixes

    • Dictation no longer gets cut off. With a quiet built-in microphone, Hex could keep only the first 2.5 seconds of a dictation and paste that as if nothing was wrong. It now keeps all of it, and a dictation that comes out empty is never dropped silently.
    • Escape no longer cancels a dictation, so a stray key press can't throw away what you said.
    • Windows: numpad keys work as shortcuts with Num Lock off, punctuation shortcuts keep working after a keyboard layout switch, and the shortcut recorder shows your layout's punctuation.
    • History: clearing History stays cleared. A background cleanup could write an old copy back, restoring cleared entries and dropping new ones.
    • Meetings: calls in Arc (including Google Meet) and WhatsApp are detected.
    • The recording indicator follows the pointer to another display during a dictation.
    • Linux: Hex runs in the background and launches at login through the desktop's Background portal, GNOME without a shortcuts portal can add Hex's shortcuts in one click, and quitting works on NVIDIA Wayland.
    • macOS: Korean and other CJK text displays correctly.

    New

    • Hebrew, with ivrit.ai's fine-tune of Whisper Large v3 Turbo.
    • Nemotron 3.5 ASR Streaming 0.6B.
    • Hex can use speech models already on your computer instead of downloading them again.
  4. v3.1.1Hex 3.1.12026-11-01 · 2026-11-01 05:43 UTC
    查看更新说明

    Hex 3.1.1

    Fixes

    • Hex opens again when a setting it can't read is in its settings file. 3.1.0 quit silently at launch for anyone who had used a recent Dev build.
    • GNOME and KDE (Wayland): asking for and approving global shortcuts works, with a safe default shortcut, a warning when it clashes with the desktop's, and a hint for asking again if the desktop declines. Paste closes its remote-control session after each paste, including on GNOME 46, and KDE restores your clipboard afterwards.
    • Linux: a long dictation no longer sits in Transcribing for a minute after the OpenCode rewrite times out. Hex's sound output closes after its tones, so it doesn't linger in the desktop's sound controls. Launch at login keeps working after Hex moves, and the window resizes properly on floating desktops.
    • Whisper Large v3 Turbo accepts Tamil and Telugu.
    • macOS: an external microphone works with the lid closed, and Settings clears stale permission rows with one-click recovery.
    • Meetings are kept for 24 hours by default; older meetings stay until you choose how long to keep them.
    • The homepage at https://anoma.ly/hex is faster.

    New in Labs

    Each is off until you turn it on in Settings › Labs:

    • Left and right modifier keys as separate shortcut keys.
    • Hold and Toggle together as a dictation shortcut mode.
    • Filters and sorting in the speech model picker.
    • Benchmark explanations in the speech model picker.
  5. v3.1.0Hex 3.1.02026-11-01 · 2026-11-01 01:39 UTC
    查看更新说明

    Hex 3.1.0

    Requirements

    • On Windows and Linux x64, Hex now needs a 2013 or newer processor with AVX2 (Intel Haswell / AMD Excavator or later). On an older processor Hex shows a message and quits instead of crashing.

    Fixes

    • Windows: the recording indicator appears reliably, and says why when it can't. Laptops with both an integrated and a dedicated GPU no longer crash at launch, and transcription survives a GPU that crashes.
    • Windows: paste works in more apps, including VS Code's terminal.
    • Meetings: starting, saving and deleting recordings, and the error banner, are fixed.
    • Annotate captures the screen and draws correctly again.
    • Linux: global shortcuts work on GNOME and KDE Wayland, including changing and cancelling them, and Hex's icon shows everywhere.
    • The recording indicator follows you: it appears on the monitor under the pointer.
    • Hex only talks to an OpenCode that's already running and no longer starts one itself.
    • Fixed a crash at startup on some Windows and Linux PCs (the CPU fix above).
    • Dictation with a Bluetooth microphone, double-tap on X11, and punctuation at the end of a dictation are fixed.

    New

    • QwenCleo-ASR 1.7B for Egyptian Arabic.
    • Recognition hints now apply to Qwen3-ASR models, and hints are merged into Vocabulary, so there's one list of names and terms.
    • Meetings can be kept for 24 hours, 7 days, 30 days, or forever.
    • New Labs toggles, each off until you turn it on: free the speech model's memory when unused, spoken formatting ("new line", "new paragraph", quotes), remove filler words, and Restore default for the Paste last dictation shortcut. On macOS also: keep the dictation on the clipboard, copy instead of paste, a second dictation shortcut, smart spacing, and subtitle export for meetings.

    Moved to Labs (now off until you turn them on)

    Turn these back on in Settings › Labs:

    • macOS: mouse buttons as shortcuts.
    • Switching the speech model with the keyboard language.
    • macOS: pausing the browser along with other media while dictating.
    • Windows: letting apps such as Discord see the dictation shortcut too.

用户点评

登录用户的真实使用记录,可按版本和硬件筛选
共 0 条

    还没有人点评。第一个说出真实体验的人最有价值。

    README 原文

    未翻译

    Hex

    Hex is local voice dictation for Mac, Windows and Linux. Hold a key, speak, and it pastes.

    How it works

    • hold the shortcut, speak, then let go
    • Hex transcribes on your device and pastes into whatever app you're in
    • the first run downloads a speech model; after that nothing leaves your machine
    • experimental features live in Labs, at the bottom of the sidebar

    Downloads

    Stable is the current release. Dev is the newest build of dev and updates itself from the dev channel.

    PlatformStableDev
    macOS (Apple silicon)DMGDMG
    Windows x64InstallerInstaller
    Windows ARM64InstallerInstaller
    Linux x86_64ExecutableExecutable
    Linux x86_64 (Debian, Ubuntu).deb.deb
    Linux x86_64 (Fedora, openSUSE).rpm.rpm
    Linux ARM64ExecutableExecutable
    Linux ARM64 (Debian, Ubuntu).deb.deb
    Linux ARM64 (Fedora, openSUSE).rpm.rpm

    Or install with:

    • Homebrew: brew install --cask anomalyco/tap/hex
    • AUR: yay -S hex-bin
    • install script (Linux): curl -fsSL https://anoma.ly/hex/install-linux.sh | sh
    • Nix: nix profile add github:anomalyco/hex

    More docs are in docs/.

    Contributing

    We don't take pull requests. Open an issue describing the bug or idea, and our agents usually turn it around fast.

    MIT licensed.