Blog Post

Anvato Launches Video ID Service

Stay on Top of Emerging Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

User uploads of copyrighted content are just a reality these days. Content owners do have some methods for protecting their IP (Viacom is still waiting on its particular choice), but none of them are simple or comfortable fixes. Most solutions, like YouTube’s, include uploading versions of all your work to a service so it can can analyze them and then find copies of them in the wild.

Maybe a bigger problem is that this is a workaround within existing intellectual property paradigms that may never escape its awkwardness. But there does seem to be some hope for another way; Lionsgate, for example, will apparently encourage fan uploads, no matter whether they are licensed, as part of its new partnership with YouTube.

Anvato, a content identification startup that had been in stealth for the last year, contacted NewTeeVee Monday for an interview alongside launching its public website this week. The Mountain View, Calif.-based company fits into the same category as existing providers Audible Magic and Vobile, but it claims its technology works better. Anvato uses picture recognition rather than audio recognition to make its matches (we’d previously road-tested Audible Magic’s audio-based tools and found them lacking). That allows it to better recognize poor quality or altered video copies.

It’s also fast and lightweight, according to Anvato CEO Alper Turgut. He said that Anvato only needs 50 to 100 KB per hour of video in order to “represent the jist of the video,” and that the service can process a million minutes of video in under 5 minutes. It doesn’t need to make a local copy of a video in order to analyze it.

Technology promises are nice and all, but we’ll be waiting to hear if Anvato signs up customers, both content owners and online video sites. Turgut promised deals are signed but didn’t name them. He said he hopes to transform copyright conflicts by giving publishers the ability to advertise on unauthorized uploads — something YouTube is already doing and other companies like Divvio hope to do. Anvato also wants to give publishers the option to replace crappy user recordings with full-quality versions, but doing that well would require deals with video hosts all over the web.

Anvato charges a monthly fee for its service as well as additional fees based on how many unauthorized videos it finds. On behalf of one customer, the company has already found a couple thousands of unauthorized videos with some 6 million plays on YouTube, according to Turgut. Content owners will pay, said Turgut, “because if the YouTube technology was working, we wouldn’t be able to find these.”

Anvato, which has just five employees, raised $550,000 in angel funding from investors that include media executives, and is looking to raise more.

4 Responses to “Anvato Launches Video ID Service”

  1. Interesting, dug into your site more and found that Alper, thanks. The main problem however still remains in that you have to get traction with the content owners in order to do all the fingerprinting. Other companies like those mentioned in the article as competitors have been trying that for a few years and it’s a major battle. I wish you luck certainly but for UGC sites, they have to sit and wait for your library to grow to the point where it is a usable service. And that could take time. I’m seeing a lot of these ID companies focusing in on helping the content owner, rather than pushing a business model that helps the UGC sites. But again, good luck to you.

  2. Josh, Anvato ContentID starts scanning approximately 30-180 minutes after the show is aired on TV. Content registration is fully automated; Anvato indexes the content (shows/episodes) live from the cable TV channel.

  3. Here’s the problem with all these services. They need the content owners to give them permission to fingerprint the original and that takes time in many different forms. Some owners are willing to do it but most are so technologically behind/backwards, even if they want to, they can’t. The other factor here is time delay. Many of these fingerprinting companies don’t get content to index until days after it airs on television. If 80% of a clips lifecycle of views happens in the first 3 days then by the time the content is first fingerprinted and copies found online it’s too late. Maybe the MSO’s should be storing identifying marks during broadcast and then providing that as a service, at least that would be closer to real-time identification? And that is going to be the biggest issue, especially as more live television, which many feel is the last big thing tv will have to offer over online, makes it’s way online. Just check out some of the rooms on,, or Yahoo Live.

  4. Jason Cheng

    Are they crawling these sites such as Youtube? They seem to have all the original shows from CBS TV. Just checked out anvato’s web site – it lets you watch the copies and the originals side by side. Wow!