What is g2-voice-commander?

This project turns Even Realities G2 smart glasses into a hands free voice commander for your own personal workflow. You tap the glasses, speak a thought out loud, and that speech is turned into a structured item in the project management tool Linear: a new issue, a full PRD document, or a comment on an existing issue. The result then shows up right on the glasses display, so you never need to pick up your phone or look at a screen. The one app included so far, voice-commander, is confirmed to work on a real device. Your speech is sent to Google's Gemini AI, which figures out what you meant and decides whether to create a Linear issue, write a PRD, add a comment, or even create a Google Calendar event if you set that up. The project describes this as one step toward a bigger idea: using the glasses as a voice input and status output surface for coordinating several AI coding tools and agents around your actual work, such as Claude Code for implementation and GitHub for pull requests. This is designed to be single user and self hosted. The code is public, but each person who uses it deploys their own private Cloudflare Worker to hold their own API keys for Gemini and Linear, and builds their own private app install. No keys are ever stored in the app itself, only in your own Worker, so nothing is shared between users and nothing sensitive ships in the installable file. Setup involves deploying the Cloudflare Worker proxy with your own API keys for Gemini and Linear, then pointing a small TypeScript app at that proxy and running it, either as a browser preview on your local network or packaged as a persistent install for the glasses. An optional Google Calendar integration needs a one time OAuth setup. The project is released under the MIT license.