keyspeake is a voice dictation and read-aloud tool for Linux. I built it because I wanted to talk to my computer the same way I could on other platforms — dictate into any application, read text back aloud, without a service in between deciding what that should cost or how it should work.
It's Linux-first, not Linux-exclusive as an afterthought: real integration with Hyprland, Sway, GNOME, KDE, and X11, not a generic wrapper that happens to run under Wayland. It'sbring-your-own-key — you pick the transcription and text-to-speech provider, you hold the API key, and your audio goes directly from your machine to that provider. There's no keyspeake server in between, and nothing about your dictation history leaves your disk unless you explicitly choose a cloud backend.
keyspeake began as a fork of y0sif's open-source project whisrs, used under its original MIT license. Its dictation core is the foundation this project was built on top of — the backends, the read-aloud media player, the GUI, and most of what keyspeake does today were added afterward, but the credit for the starting point belongs there.
I maintain keyspeake myself. If something's broken or missing, email me — it goes to a real person who can actually fix it.