Komi Extension
Voice DevOps.
An always-on, wake-word activated DevOps companion for Git & GitHub operations, connecting natural speech controls directly to code editors.
The Problem
Context-switching between writing code, running Git terminal commands, creating repositories, and verifying remote push statuses breaks developer concentration and increases cognitive load.
The Solution
Komi Extension embeds a standalone voice automation capture loop inside a VS Code editor extension. Utilizing wake-word listening, transcription models (Google Gemini), and schema-enforced JSON action contracts, the extension translates spoken intents (e.g. "push my changes", "create repository") into shell operations and API calls without requiring terminal interaction.
Core Features
Wake-Word & Passive Voice Loop
Monitors passive audio inputs using a wake word timer threshold, keeping a high-sensitivity action window open for commands.
Gemini Schema-Enforced Reasoning
Parses transcripts using structured JSON tool-call schema rules to eliminate incorrect action executions.
Git & GitHub API Automation
Automates workspace repository initialization, file staging, commit sequences, and GitHub API repository creation.
VS Code Panel & Status Bar UI
Provides interactive buttons, masked environment Secret indicators, status indicators, and tree views inside the host editor.