InkSpoke is a cross-platform desktop Voice OS that turns speech into formatted, context-aware text and types it into whatever app you are already using. It runs offline-first, with optional cloud assistance.
This repository hosts the public releases for the InkSpoke desktop app. Source code lives in a separate private repository.
Get the latest version for your platform from the Releases page, or from inkspoke.app/download.
| Platform | Requirements |
|---|---|
| macOS (Apple Silicon) | macOS 13 Ventura or later, Apple M1 or newer |
| Windows (x64) | Windows 10 (1903) or later. Optional NVIDIA GPU for CUDA acceleration |
| Linux (x64) | Ubuntu 22.04, Fedora 38, or equivalent. PulseAudio or PipeWire |
Every release is also mirrored to a Cloudflare-backed mirror. Replace the GitHub download host with:
https://dl.inkspoke.app/releases/
The app updates itself automatically and will fall back to this mirror when GitHub is unreachable.
- Website — https://inkspoke.app
- Documentation — https://docs.inkspoke.app
- Issues and feature requests — please use the Issues tab
- Email — hello@inkspoke.app
InkSpoke is proprietary software. © PMT Labs LLC. All rights reserved.
Third-party components are listed in the notices bundled with each release.