Face.com

» see face.com as it lived, on the wayback machine

~ the goodbye ~

face.com never had a consumer surface i could touch. it was an api. developers wired it into photo apps to auto-tag faces, and the biggest visible deployment of that api was the two facebook add-ons face.com itself shipped, photo finder and photo tagger. from ankara in 2011 i knew face.com only as the vendor name that showed up in api reference pages when i went looking at how facebook tagging worked under the hood.

the company was a tel aviv startup founded in 2009, part of the same israeli computer-vision cluster that produced a decade of biometric plays sold to us tech buyers. by february 2011 face.com claimed 18 billion faces scanned across its api and its two facebook apps. by mid-2012 around 10,000 developers were pulling from the endpoint. the total funding was small by the standards of tech, about 5.3 million dollars across the series a and series b, which is what a face-detection api looked like before the recognition sector consolidated.

face.com's product answered a question facebook had been trying to solve internally, which was how to tag every face in every uploaded photo without a human clicking on it. the answer was a small tel aviv company's api, and that answer was worth around 55 to 60 million dollars to facebook by the time the offer went out.

~ the chronicle ~

~ the acquisition ~

on 18 june 2012 facebook announced the acquisition. the deal number was not disclosed officially, but the range that surfaced in press was 55 to 60 million dollars. the same day, face.com posted a note that the public api would shut down within one month. the deprecation window was that short because facebook did not want face.com's tech running under any other company's product. the plain-language description of the deal is acqui-kill, buy the tech and take the team into the acquirer's stack.

the crew that moved to menlo park included the founders and most of the israeli engineering side. their work fed into what facebook later shipped as deepface, the internal facial-recognition system that ran on billions of facebook photos through the mid-2010s. deepface was the successor product in the plain lineage sense, same code lineage and same people, just relocated inside facebook.

~ how it actually ended ~

face.com's public api was gone by mid-july 2012. the developer forum posts from that window were mostly people asking for migration paths that did not exist. some moved to microsoft's face api on azure once that launched later, others rewrote around opencv or dropped the tagging feature entirely. every third-party photo app that had been running on face.com's endpoint had four weeks to unplug.

the face.com domain sold to an unrelated e-commerce operator in december 2015, which is the standard afterlife for a name that used to point at a working product. the wikipedia article still lists face.com's peak numbers, 18 billion photos scanned across around 10,000 developers, but the api documentation itself has been gone for fourteen years now (in 2026).

~ what got buried with it ~

the acqui-kill pattern got a name in the tech press around this window, and face.com was one of the entries that made the name stick. buy a small company for a working piece of tech, quietly move the team into the acquirer's stack while closing the outside product. facebook did this with sparrow and gowalla and glancee and beluga in the same 2011 to 2013 stretch. the ones i noticed at the time were the consumer apps because they had users who felt the shutdown. face.com was a b2b endpoint, so what remained after the shutdown was mostly developer forum threads and not much else. still counts as a death.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »