Clarifai is an artificial intelligence startup with $10 million in the bank and a small cadre of well-trained researchers. Its main schtick is an API that combs through loads of visual and video data, then packages it up neatly. Its clients include Unilever, video site Vimeo and BuzzFeed.
But on Wednesday, the startup is putting out its first consumer product: An app, called Forevery, that stores your photos then uses machine learning to organize them with granular detail. It automatically generates tags, mining from a database of some 11,000 terms both concrete (tree, dogs, cake) and quite abstract (idyllic, togetherness, love).
Sound familiar? It should. That’s a lot like Google Photos, which the search giant rolled out in May and quickly expanded to 100 million monthly users. Matthew Zeiler, Clarifai’s CEO and founder, is undeterred by that number. “This is much more in-depth and personalized than you get out of Google Photos,” he said about Forevery.
Clarifai’s release is another example of how deep learning, the all-the-rage machine intelligence method, is squeezing into more consumer products.
It’s also a sign of how treacherous the path is for startups in this field. While Clarifai would like its consumer app to catch on, its main purpose is absorbing more data for its primary offering for enterprises. It just so happens that, last week, Google released a cloud tool that lets developers tap its image recognition capabilities — an API akin to Clarifai’s product.
Here’s what Clarifai, and many others facing the Internet behemoth, are banking on: For Google, this is one project among many. For Clarifai, “It’s our focus,” Zeiler said.
The CEO, a young PhD who nabbed a key deep learning prize in 2013, spent a brief spell as a researcher at Google. Both his CTO and the chief engineer behind Forevery, Keith Ito, spent around a decade each there. Among the startup’s investors: Google Ventures.
Forevery is available just for iOS to start.
This article originally appeared on Recode.net.