// AI NEWS · FUNDING

Clipto Hits a $250M Valuation for On-Device AI Search Across Your Video Archive

Announced August 31, 2026, the three-year-old media-search startup raised $15M and now searches videos, audio, images, and files locally on your device — and can hand indexed results to ChatGPT or Claude over the Model Context Protocol.

2026-08-31 · by Moe Ameen

What happened

Clipto, an AI media-search startup, has raised $15 million in an all-equity round at a $250 million post-money valuation, TechCrunch reported on August 31, 2026. The San Francisco company — with additional teams in Singapore and Hong Kong and just over 20 employees — is roughly three years old and led by founder and CEO Henry Kang. Investors in the round include HSG (formerly Sequoia China), GL Ventures, EnvisionX Capital, Palm Drive Capital, 522 Ventures, and angels Hans Tung and Lu Zhang.

What Clipto does is index the media already sitting on your computer — videos, audio, images, meetings, and other files — and let you search it in natural language. You describe the moment you're looking for and it retrieves the file or the clip. It recently added support for the Model Context Protocol (MCP), so an AI tool like ChatGPT or Claude can query that index and pull selected results into a chat, though the company says access requires the user's active request and authorization each time.

The differentiator Clipto leans on is that all of this processing runs locally on the user's device rather than in the cloud — a privacy stance that matters when the corpus is your own raw footage, recorded calls, and personal photos. The company reports more than $15 million in annual recurring revenue, says it remains profitable on a net-income basis, and states that more than 30 million people have used its products since launch, with hundreds of thousands of paying subscribers.

The important scope note for creators: Clipto is a retrieval layer, not a production or publishing tool. It finds the moment in your archive. It does not cut it into a short, caption it, brand it, or post it — turning a located clip into finished content that ships is a separate job.

Why it matters for creators

  • Most creators are sitting on terabytes of unwatched footage. A natural-language search that surfaces the exact moment — "the part where I explained pricing" — turns a dead archive into a usable content library.
  • On-device processing is the real hook. Your raw footage, recorded calls, and photos never leave the machine, which lowers the trust cost of indexing sensitive or unreleased material.
  • The MCP support means an AI assistant can now reach into your own media, not just the public web — but the connector reads and retrieves; it does not produce or publish anything.
  • Finding the clip is only step one. The value a creator captures depends on how fast that located moment becomes a captioned, reframed, on-brand post across every platform they publish to.
  • A $250M valuation on a retrieval tool signals real demand for taming media archives — which sharpens, not solves, the downstream question of what you make from what you find.

How to act on this with Kompozy

Clipto answers "where is that moment in my footage?" It's a memory layer for your own archive, and a good one — you describe the clip and it hands you the file. What it deliberately doesn't do is anything after retrieval: no cutting, no captions, no branding, no posting. That's the exact seam [Kompozy](/) is built to pick up, and the pairing is clean because the two tools never overlap. Use Clipto to locate the golden 90 seconds buried in a two-hour livestream or an old webinar recording; then drop that long-form clip into Kompozy and let [Clipped Shorts](/glossary/output-buckets) cut it into vertical shorts, burn in branded captions, and reframe it to 9:16, 1:1, and 16:9 — the production step Clipto's search leaves untouched.

From there Kompozy expands one found moment into a week of content. A single retrieved clip becomes the raw material for [Persona Shorts](/glossary/persona-shorts) and HeyGen avatar video that reframe the take on camera, a brand-exact [Carousel](/glossary/hyperframes), Quote Graphics, a Blog Article, and an Email Newsletter — every piece written in your voice under the [Persona Brief](/glossary/persona-brief) so a batch mined from old footage still reads as one brand. Then [Autopilot](/glossary/autopilot) schedules and publishes it across the eight social platforms plus blog and email behind a per-post review. Clipto turns your archive into a searchable library; Kompozy turns what you pull from that library into finished, on-brand posts that actually ship — retrieval and production, doing separate halves of the same job.

Quick takeaways

  • Clipto raised $15M at a $250M post-money valuation, reported August 31, 2026; founder/CEO Henry Kang, HQ in San Francisco, backed by HSG, GL Ventures, EnvisionX Capital, Palm Drive Capital, 522 Ventures, plus Hans Tung and Lu Zhang.
  • The product indexes videos, audio, images, and files on your device and searches them in natural language, with all processing kept local and recent MCP support to feed results to ChatGPT or Claude.
  • Clipto reports 30M+ users since launch, hundreds of thousands of paying subscribers, more than $15M ARR, and profitability on a net-income basis.
  • Clipto is a retrieval layer, not a content engine — it finds the moment; Kompozy is what cuts, brands, expands, and publishes it across nine platforms.

Frequently asked questions

What is Clipto and what did it announce?

Clipto is an on-device AI media-search startup that indexes videos, audio, images, meetings, and files on your computer and lets you search them in natural language. On August 31, 2026 it announced a $15M all-equity round at a $250M post-money valuation, led by founder and CEO Henry Kang, with all processing running locally on the device.

Does Clipto edit or publish video?

No. Clipto is a retrieval and search tool — it finds the moment in your archive and can hand results to AI tools like ChatGPT or Claude over the Model Context Protocol. It does not cut clips, add captions, brand content, or publish to social platforms. Turning a located clip into a finished, published post is a separate job.

How does Clipto keep data private?

Clipto states that its indexing and search run locally on the user's device rather than in the cloud, so your footage, recorded calls, and photos stay on the machine. When an AI tool queries the index via MCP, the company says access requires the user's active request and authorization each time.

How do creators turn Clipto search results into content?

Use Clipto to locate the exact moment in your footage, then feed that clip into a generation-and-publishing engine like Kompozy. Kompozy cuts it into captioned vertical shorts, reframes it per platform, and fans it into avatar video, carousels, images, a blog, and a newsletter in your Persona Brief voice — then schedules and publishes across the eight social platforms plus blog and email.

Related news

← All AI news · Get started →