Numen / Blog

What, Why and Where are we going?

Numen is a project that enables people to use a computer with their voice. The main motivation is to help people with strain or limited use of their hands.

I wrote numen because I found nothing simple that just worked, everything was a framework to configure with application specific rules and grammars, and was trying to be your text editor and voice assistant, layered on top of something proprietary. I just wanted an efficient keyboard alternative that worked everywhere without hassle.

With numen you can simply type by saying syllables and literal words, and can be really efficient using standard tools. It's great writing and manipulating code in Vim and it's comfy surfing the web on qutebrowser, and you can just use the phrases you use everywhere and you don't need to configure anything.

The other approaches I've seen put needless complexity into the voice control and then only work with specific applications and graphical environments. Numen works universally in X11, Wayland and TTYs, and will soon be able to turn a Raspberry Pi into a voice input device for any computer.

I think it will be great when numen is just there in your package manager, ready to install, and you're off with a look at the manpage.

At one point I thought I'd need something like eye tracking, but there's always been a better way by voice. I really don't think dragging application windows with your eyes or using it to select text is ideal. The computing environment is so integral to voice input that I decided to package my own including my tiling window manager setup and programs.

I'm very thankful for the recent leaps in free and open source speech recognition, a couple years ago keyword recognition wasn't accurate enough and now I'm using literal transcription to type this! I hope numen can be the leap for free and open-source voice control.