SKI
A cross-platform desktop app that adds a voice interface to coding agents such as Claude Code, Cursor, Codex and Gemini CLI: you talk to the agent out loud and hear its answers in a natural voice, with everything running on the local machine.
SKI lets a developer talk to a coding agent instead of typing, and hear it answer back in a synthesized voice; the vendor states nothing said is sent anywhere beyond the local machine. This repository hosts only the compiled release binaries and the auto-update feed for macOS, Windows and Linux - the application's source code lives elsewhere.
On macOS the build is notarized and updates download in the background and apply on relaunch; on Windows and Linux updates are applied manually from a Preferences panel. Every release is cryptographically signed, and the README gives the exact shasum, PowerShell or sha256sum command to verify a downloaded installer against the hash published in that release's notes. First launch on Linux downloads a roughly 620 MB speech-recognition model, which is checksum-verified and resumable if interrupted; the voice output model ships bundled and needs no separate download.
Supported platforms are macOS 14.4 or later on Apple Silicon, Windows 10 or 11, and Ubuntu 24.04 or Debian 13 or newer on Linux.