Vercept's new AI agent controls your computer like a human
Vercept just launched Vy, an AI agent that runs natively on your computer, seeing your screen and directly using your signed-in apps. It is powered by VyUI, a new model that beats OpenAI, Google, and Anthropic on UI grounding benchmarks, and is available to download today.
Taking the wheel on your desktop
Vercept just launched Vy, a native AI agent designed to see and interact directly with your computer. It is available for download starting today.
Unlike cloud-locked chatbots, Vy runs natively on your machine. That means it can directly tap into the desktop apps and accounts you are already logged into, operating your software just like a human would.
Powering the agent is VyUI, a dedicated model built to bridge the gap between text instructions and graphical user interfaces. Vercept claims VyUI sets a new state-of-the-art on UI grounding benchmarks like ScreenSpot v1, ScreenSpot v2, and GroundUI Web, outperforming frontier models from OpenAI, Google, and Anthropic.
Why it matters
Desktop computer-use agents are quickly becoming the next battleground for productivity. We think Vy’s native approach is clever because it bypasses the friction of API configurations and account authentications; it simply takes over the workspace you already have open.
However, handing an AI direct control of an unlocked computer comes with glaring risks. If an agent misclicks an action or misinterprets visual context, the fallout is far worse than a standard text hallucination. Still, if Vercept can genuinely beat the tech giants on UI grounding benchmarks, it proves that specialized screen models might carve out a viable edge over massive general-purpose models.
Sources (1)
- @vercept_ai(opens in a new tab)Originalx.com