Tanush Shah

FlowPilot is a voice-native operating layer for Windows that turns human intent into real desktop actions. Most AI tools stop at conversation: they answer, explain, or generate text. FlowPilot goes further by helping users actually do the thing. With global hotkeys, users can instantly switch between raw dictation, AI-polished writing, and a Jarvis-style assistant that understands spoken intent and performs safe computer actions. The problem is simple: modern computers constantly break flow. We waste time switching tabs, finding files, rewriting text, opening tools, copying information, and remembering shortcuts. FlowPilot removes that friction by letting users say what they want in plain language and letting the system help execute it. The app combines local-first speech recognition, smart writing, a top-center voice orb, selected-text context, memory, workflows, browser/file/app actions, and confirmation gates for sensitive operations. It can type literal speech, clean messy spoken thoughts into ready-to-send writing, open apps, manage browser tabs, prepare Gmail drafts for review, search local files, remember user-approved routines, and assist with what is selected or visible on screen. For the Cursor track, FlowPilot reimagines everyday computer interaction around speed, flow, and human expression. The LLM is not just a chatbot here — it is the reasoning layer between intent and execution. The long-term vision is a computer interface where the keyboard becomes the trigger, not the main input method, and users stay in control without losing momentum.