Hark, an AI startup founded this year by roboticist and serial entrepreneur Brett Adcock, launched Handoff, a computer use agent that autonomously performs web-based tasks on behalf of users. The system navigates websites without human intervention, handling actions like ordering food on DoorDash, booking flights on United and Delta, and messaging job candidates on LinkedIn.
Computer use agents represent an emerging category of AI systems trained to interact with digital interfaces the way humans do, clicking buttons, filling forms, and navigating menus. Handoff joins competitors like Anthropic's Claude and OpenAI's tools in this space, but Hark positions its agent as among the fastest and most affordable options available.
The startup kept a low profile before today's announcement, operating as what insiders call a "secretive" venture. Adcock's background in robotics and entrepreneurship signals Hark's approach differs from purely software-focused competitors. The company is betting that autonomous agents handling routine digital tasks represent genuine commercial value.
Public sign-ups launched today, with broader availability expected later this month. Pricing and performance benchmarks against competitors remain unclear from the announcement, though Hark emphasizes affordability and speed as differentiators. These claims matter because computer use agents currently struggle with reliability, complexity, and cost when handling multi-step tasks requiring real reasoning.
The timing reflects accelerating competition in autonomous AI agents. Both large labs like Anthropic and newer entrants recognize this space as potentially lucrative. If Handoff delivers on speed and cost promises, it could capture users frustrated by expensive API calls or slow response times from existing solutions.
Success requires handling edge cases, security concerns, and reliability across diverse websites using different layouts and structures. Hark's robotics heritage may provide advantages in systematic problem-solving, but execution matters more than pedigree. The market's appetite for AI agents depends