OneSay is turning voice from an input method into an everyday way to operate: understanding what you want to express, understand, change, or do next; preserving the context of the app and content in front of you; invoking the right capability; returning the result to the right place; and gradually adapting to how you communicate and think—all within clear permission boundaries.
Software is organized around features. Life is not.
People do not first classify a thought as work, life, learning, or communication before deciding how to express it. Messages, emails, notes, questions, and plans all come from the same person. OneSay should be built around that person—not around a single task or category.
Where we’re going
Voice should not be confined to a single text box or category of work. It can carry the intent to write, translate, edit, ask, learn, plan, decide, and act. OneSay begins today with four core language intents on desktop. Over time, it will connect more applications, knowledge, and actions—so people do not need to find the right tool before expressing what they want to say, understand, or do next.
Our principles
OneSay should shorten the distance between intent and result—not cross the final line on the user’s behalf. Writing back should require confirmation. External search should be optional. Diagnostics should be opt-in. Sources should remain visible. Memories should be reviewable and removable. Product and privacy claims must always reflect the current implementation. The more a system understands a person, the more carefully its permissions, responsibilities, and ability to forget must be designed.
Open roles are published only after they are confirmed. No placeholder positions. No manufactured hiring activity.
View role updates