See it in action
https://github.com/user-attachments/assets/ed91275b-9716-4aab-a41a-eb5b3807abaf
Press play, then use the speaker control to turn on sound.
Download the 45-second promo (MP4) · Video details
An animated macOS workflow showing dictation into sample writing apps, with voiceover and music. Timing is illustrative; Linux uses Copy and manual paste.
Actual app recording: model selection, local recognition, and Copy.
Download
Textify 0.2.0 Preview 23
The Mac app is Developer ID signed and notarized by Apple. Windows and Linux installers are unsigned and have not been tested on physical hardware; their automated installer and launch checks pass. Second-Mac fresh-install and update checks remain outstanding. This is a preview release.
A hardware GPU is required; CPU-only inference and Intel Macs are not supported. Speech models download separately. Updates are installed manually from GitHub Releases.
Installation instructions · SHA-256 checksums · Platform QA
Get started
- Install Textify. Download the installer for your platform. On macOS, drag Textify into Applications and grant Microphone and Accessibility access when prompted.
- Choose a model. Open Transcription models and download a supported version. After verification, click Use model when shown, then choose a supported dictation language.
- Start dictating. Enable the global trigger, focus a text field, hold your shortcut, speak, then release. You can also use the in-app microphone button and choose Copy.
| Desktop | Default shortcut | Your text |
|---|
| macOS | Right Command | Pasted into the original app when it is still focused |
| Windows | Right Control | Pasted into the original app when it is still focused |
| Linux X11 | Right Control | Choose Copy, then paste manually |
| Linux Wayland | Chosen through the desktop's GlobalShortcuts portal | Choose Copy, then paste manually |
On Wayland, shortcut availability depends on the desktop portal. The microphone button remains available when a global shortcut cannot be enabled.
Built for everyday dictation
Models
Whisper is available on macOS, Windows, and Linux. Apple Silicon also supports Parakeet, Qwen3-ASR, and Confucius4-R2T2. Model and language availability varies by platform and catalog entry; the app shows the supported languages for each choice.
Model binaries are downloaded separately from pinned sources and checked against signed metadata before use. See the desktop model and runtime details for the supported versions.
Privacy
Audio and dictated text are processed locally and held in memory for the active session. Textify has no transcript history, accounts, analytics service, or speech upload. Activity stores only daily numeric totals on your device.
Model downloads are explicit network requests. Their hosts receive ordinary request metadata, such as an IP address and the requested model file. Dictation itself works offline after setup.
Automatic paste restores the previous clipboard when it has not changed; explicit Copy replaces the clipboard normally. Third-party clipboard managers may retain copied text. Learn more about desktop behavior and clipboard handling.
Development
The current desktop app lives in electron/. Start with the desktop development guide for platform prerequisites, local builds, and verification. Release packaging is documented in the desktop release guide.
Support and contributions
Bug reports, feature requests, and model suggestions are welcome. Contributions currently happen through GitHub Issues; external pull requests and model binary submissions are not accepted. See the contribution policy.
Please report vulnerabilities privately through the process in SECURITY.md.
License and acknowledgments
Textify is licensed under the Apache License 2.0. It builds on open-source speech runtimes and models from many communities.
Acknowledgments · Third-party notices · Third-party licenses
