The generative AI boom solved one problem and created another. Producing content is trivial now; finding any of it later is not. Clipto, a San Francisco startup with teams in Singapore and Hong Kong, just raised $15 million in an all-equity round at a $250 million post-money valuation to attack that second problem. Investors include HSG (formerly Sequoia China), GL Ventures, EnvisionX Capital, and Palm Drive Capital.
Clipto indexes videos, audio, images, meetings, and documents on a user’s computer, then lets them search by describing what they need — or by asking ChatGPT, Claude, or another AI tool to find it. Founder Henry Kang has been circling this idea for two decades. As a Carnegie Mellon PhD student in 2006, he worked on robots that recorded their surroundings and remembered where objects were. He later applied the concept to a closet-organizing app, then sold his video startup ZenVideo to Tencent in 2020. “We don’t have a content shortage,” Kang told TechCrunch. “The opposite is true. We have too much content.”
The company claims more than 30 million users since launch, hundreds of thousands of paying subscribers, and $15 million in annual recurring revenue as of early 2026 — while staying profitable with just over 20 employees. Two weeks ago, Clipto added Model Context Protocol support, letting AI assistants query indexed files with explicit user authorization. Processing runs locally on-device, which Kang frames as both a privacy and performance advantage.
The obvious question is whether this becomes a standalone category or gets absorbed into existing products. Adobe, Apple, and Google already offer natural-language search within their own ecosystems. Clipto’s pitch is broader: search across everything on your machine, regardless of where files live. That’s a real differentiator, but it also puts the startup in the path of companies with far deeper pockets and pre-installed distribution. The $15 million will go toward on-device AI models and agent integrations — a bet that consumers will pay for search that actually understands what’s on their own computers.
