Clipto's on-device AI search hits $250M valuation: what it means for local-file AI tools

Clipto, a three-year-old startup that indexes video, audio, images, and documents on your local machine and lets you search them with natural language or AI assistants, just raised $15 million at a $250 million valuation. The company is already profitable with $15 million in ARR and over 30 million users. For builders, the interesting part isn't just the valuation: it's that Clipto runs everything on-device, added Model Context Protocol (MCP) support, and is betting that local AI search is a standalone product, not just a feature of Adobe or Google.

On-device indexing across media types

Clipto indexes videos, audio, images, meetings, and other files on a user's computer. Instead of scrolling through folders, users can search by describing what they're looking for or ask ChatGPT, Claude, or other AI tools to find it for them. All processing runs locally on the user's device without requiring cloud services, according to founder Henry Kang.

The product was initially built for video creators managing footage across drives, but creators now make up only about one-quarter to one-third of users. The rest include lawyers, doctors, researchers, marketers, HR professionals, professors, and students - anyone who needs to search across large local media libraries.

MCP integration opens local files to AI agents

About two weeks ago, Clipto added support for Model Context Protocol (MCP), a standard that allows AI applications to connect with outside sources of information. Access to indexed files requires the user's active request and authorization, and AI applications can only retrieve information within the scope the user specifies. This is a practical move for builders: it means you can build agents that query local media through a standard protocol without sending data to the cloud.

Competing with incumbents on cross-media search

Adobe already offers AI-powered search in Premiere, while Apple Photos and Google Photos allow natural-language search within their own ecosystems. Kang argues that Clipto can search across videos, audio, images, and documents, while the larger players focus on files stored within their own services. Clipto also lets users make selected information available to AI tools like ChatGPT and Claude. The new funding will go toward the AI models and computing infrastructure needed to run Clipto on consumer hardware, as well as integrations with more AI agents.

Caveats

The company declined to provide a precise subscriber count or average revenue per customer, though Kang said a significant portion of customers remain subscribed for more than two years. The $250 million valuation is high relative to $15 million ARR (about 16.7x), but the company is profitable on a net-income basis. The on-device processing claim is central to the product's value proposition, but independent performance benchmarks at terabyte scale are not yet available. The market also includes well-funded incumbents with native search features.