Shadow

Shadow is a desktop application designed to function as an AI-powered assistant that integrates deeply into a user’s workflow. It operates primarily through two modes: automatic participation in online meetings, and on-demand activation via keyboard shortcuts across any other application. In meetings, Shadow launches automatically when a video call begins, transcribes every spoken word in real time, and detects shared screens to capture screenshots — ensuring that no detail is missed. After the meeting, it can generate structured meeting notes, action items, and summaries based on the captured audio and visual content, allowing users to review discussions without replaying recordings. Outside of meetings, users can press a predefined keyboard shortcut at any moment to trigger Shadow’s capture function. This captures whatever is visible on the screen (e.g., a document, a webpage, a code editor) and whatever the user is saying aloud. Shadow then processes this input through user-defined “Skills” — customizable prompts or templates that transform the raw captured data into specific outputs. For example, a user reading a long article can capture a paragraph and ask Shadow to summarize it, rewrite it as an email, or translate it into code. A developer reviewing a bug report can capture the screen and voice notes, and Shadow can generate a formatted bug description with steps to reproduce. For customer support representatives, Shadow can listen to a client call, capture key details from the screen, and produce a draft reply email with appropriate tone and context. The application also supports voice typing: users can dictate notes, messages, or documents through Shadow, which transcribes speech into text with high accuracy. Quick replies — generating short responses to messag