VOICE → WORDS https://hex.kitlangton.com
  • Swift 99.1%
  • TypeScript 0.8%
  • Ruby 0.1%
Find a file
History 2025-11-14 15:31:04 -05:00
.cursor/rules init 2025-02-21 19:27:22 -06:00
.github Revert to local-only release workflow 2025-11-14 14:48:35 -05:00
bin Add Sparkle generate_appcast binary 2025-11-14 14:16:46 -05:00
docs Update docs and tools 2025-11-14 13:21:01 -05:00
Hex Add text transformation pipeline with drag-to-reorder UI 2025-11-14 15:25:49 -05:00
Hex.xcodeproj Bump version to 0.3.0 2025-11-14 13:44:30 -05:00
HexCore Add text transformation pipeline with drag-to-reorder UI 2025-11-14 15:25:49 -05:00
.gitignore Revert to local-only release workflow 2025-11-14 14:48:35 -05:00
.swiftlint.yml add github ci steps for releasing and building 2025-05-30 09:35:57 -05:00
AGENTS.md Add configurable hotkey for pasting last transcript, download cancellation, and model deletion 2025-11-13 22:13:08 -05:00
CLAUDE.md Add 0.3s safety threshold for modifier-only hotkeys 2025-11-14 10:24:02 -05:00
hex.rb Add TypeScript release pipeline with tag-based versioning 2025-11-14 12:46:43 -05:00
LICENSE Add MIT license and minor UI/model adjustments (resolves #84) 2025-11-13 22:15:00 -05:00
Localizable.xcstrings Update models and localization strings 2025-11-14 13:20:56 -05:00
README.md Update README: remove Discord link, highlight Parakeet as default transcription engine 2025-11-14 15:31:04 -05:00

Hex — Voice → Text

Press-and-hold a hotkey to transcribe your voice and paste the result wherever you're typing.

Download Hex for macOS

Note: Hex is currently only available for Apple Silicon Macs.

I've opened-sourced the project in the hopes that others will find it useful! Hex supports both Parakeet TDT v3 via the awesome FluidAudio (the default—it's frickin' unbelievable: fast, multilingual, and cloud-optimized) and the awesome WhisperKit for on-device transcription. We use the incredible Swift Composable Architecture for structuring the app. Please open issues with any questions or feedback! ❤️

Contributing

Issue reports are welcome! If you encounter bugs or have feature requests, please open an issue.

Note on Pull Requests: At this stage, I'm not actively reviewing code contributions for significant features or core logic changes. The project is evolving rapidly and it's easier for me to work directly from issue reports. Bug fixes and documentation improvements are still appreciated, but please open an issue first to discuss before investing time in a large PR. Thanks for understanding!

Instructions

Once you open Hex, you'll need to grant it microphone and accessibility permissions—so it can record your voice and paste the transcribed text into any application, respectively.

Once you've configured a global hotkey, there are two recording modes:

  1. Press-and-hold the hotkey to begin recording, say whatever you want, and then release the hotkey to start the transcription process.
  2. Double-tap the hotkey to lock recording, say whatever you want, and then tap the hotkey once more to start the transcription process.

💡 Note: The first time you run Hex, it will auto-download the Parakeet model in the background. This is incredibly fast and optimized for the cloud—you'll be up and running in no time!

License

This project is licensed under the MIT License. See LICENSE for details.