Clipto has raised $15 million in an all-equity round that values the three-year-old AI media-search startup at $250 million after the investment. The company is betting that the flood of files created and collected with generative AI will make personal search a larger problem than any single application can solve.
The San Francisco-headquartered company indexes video, audio, images, meetings and other files stored locally. Users can describe what they want to find or connect the index to ChatGPT, Claude and other AI systems. Clipto says this cross-format approach separates it from search features tied to individual products such as Adobe Premiere, Apple Photos or Google Photos.
HSG, formerly known as Sequoia China, and GL Ventures participated alongside EnvisionX, Palm Drive, Hans Tung, Lu Zhang and 522 Ventures. Clipto also operates teams in Singapore and Hong Kong.
A search layer across local media
Founder Henry Kang’s interest in media retrieval dates to work at Carnegie Mellon University in 2006, where robots recorded their surroundings and needed ways to recall objects. He later created video startup ZenVideo, which Tencent acquired in 2020, before founding Clipto with members of his previous team in 2023.
Clipto first targeted video creators, but the company says that group now represents only 25% to 33% of its users. Its customer base has expanded to lawyers, doctors, researchers, marketers, human-resources workers, professors and students—people whose source material may span recordings, meetings, documents and images.
The startup claims 30 million users and hundreds of thousands of paying customers. It declined to disclose an exact subscriber count or average revenue per customer, limiting outside assessment of those figures. Kang said a significant portion of subscribers have paid for more than two years.
Clipto reached $15 million in annual recurring revenue at the beginning of 2026 and is profitable on a net-income basis, according to the company. It has achieved that with a staff of just over 20 people, a compact operation for the valuation attached to the new round.
Agents become another route into the index
About two weeks ago, Clipto added support for the Model Context Protocol, or MCP. The integration lets AI agents retrieve files from the index when a user makes a request and grants authorization.
Kang said access is limited to the scope of the request and that file processing occurs locally. Those are company descriptions of the system; TechCrunch’s report does not provide an independent security or privacy audit. For users indexing sensitive legal, medical or workplace material, the distinction between a claimed local process and verified controls will be important.
The agent connection expands Clipto’s role beyond a standalone search box. A user could ask an AI system for information contained in personal media, with Clipto acting as the retrieval layer behind the interaction. That makes the index useful to multiple assistants instead of requiring every assistant to build its own file catalog.
It also creates a technical challenge. Search must recognize content across different file types while returning only the material a user intended to expose. As agent integrations multiply, authorization boundaries may become as significant as search relevance.
Funding targets models and consumer hardware
Clipto plans to use the funding to develop models that can run on consumer devices and to deepen integrations with AI agents. Keeping more work on local hardware could support the company’s privacy proposition while reducing reliance on remote compute, although the company has not disclosed performance targets or a delivery timeline.
The round reflects a broader commercial premise in Clipto’s pitch: generative AI can increase the volume of content faster than people can organize it. If that continues, tools that retrieve existing material may become as valuable as tools that produce new material.
Competition is substantial because major software and photo platforms already offer AI-assisted search. Clipto’s opportunity depends on whether users value one index that crosses service and file boundaries enough to adopt another layer of software.
Its reported revenue, profitability and paid retention give the company more evidence than a purely speculative search product. The undisclosed subscriber economics and unverified privacy controls still leave material questions. The new capital will test whether a small team can turn cross-media retrieval into infrastructure for the agents now competing to understand users’ personal files.
