Clarifai

Clarifai Inc. is an artificial intelligence company founded in 2013 by Matthew Zeiler that built a deep learning platform for modeling unstructured image, video, text and audio data. Its original product was an image and video recognition API for developers and businesses, later expanded into a full AI lifecycle platform covering data labeling, custom model training, workflow orchestration and deployment. Platform capabilities have included object classification, detection and tracking, geolocation, visual search, content moderation, OCR, face/person detection, natural language processing and audio recognition, with pre-built models (for example general concept, food/travel and NSFW models) alongside custom model training. Deployment options spanned cloud, on-premise, bare-metal and classified environments, plus edge deployment on hardware ranging from server systems to camera sensors and drones.

Products launched after the company's 2016 Series B included Spacetime Search, Scribe automated data labeling, and Flare Edge AI; the company also positioned an "AI Lake" concept for managing inputs, concepts, models, workflows and applications in one place, with access governance. Clarifai released version 2 of its API in 2016 adding custom training and visual search, a mobile SDK in 2017 enabling offline operation, and an on-premise solution in 2018. Early on it also offered a consumer iPhone app called Forevery.

In more recent years the company's work centered on AI inference and compute orchestration. In May 2026, Nebius (NASDAQ: NBIS) announced that Clarifai's core engineering and research team, led by Zeiler, was joining Nebius, that Nebius would acquire Clarifai's patent portfolio covering AI inference and compute orchestration, and that it would receive a non-exclusive, perpetual license to Clarifai's modern AI inference and reasoning technology stack. The scope explicitly excluded Clarifai's legacy computer vision models and any IP, products, services or commercial arrangements tied to Clarifai's US government and defense programs.

Founding story

Matthew Zeiler grew up in a small town in Canada and studied computer neural networking under Geoffrey Hinton at the University of Toronto, then earned a PhD in machine learning and image recognition at New York University supervised by Rob Fergus. He interned at Google with Jeff Dean on the Google Brain deep learning project and received job offers from Google, Microsoft and Facebook, but left the internship early and returned to New York to found Clarifai in 2013, stating that the technology available to him in New York was more advanced than what he had been working on in California. That same year he placed in the top five spots in image classification at the 2013 ImageNet Challenge.

Business model

Clarifai sold access to its AI platform to developers and organizations, offering free and paid tiers of its recognition APIs, custom model training, data labeling (Scribe LabelForce) and workflow tooling, with deployment across cloud, on-premise, bare-metal, edge and classified environments for enterprise and public sector buyers.

Paid access to API-based recognition services, custom training and platform products, sold to enterprise and public sector customers alongside free tiers; the company reported more than doubling revenue in 2020.

Traction

The company reported more than 130,000 global users of its platform in 2021 and said it more than doubled revenue in 2020. It had about 100 customers in 2017, 33 employees in 2015, more than 100 employees in 2021 and 120+ employees per a later company profile.

Latest developments

On 12 May 2026 Nebius announced that Clarifai's core engineering and research team, led by founder and CEO Matthew Zeiler, is joining Nebius, that Nebius will acquire Clarifai's patent portfolio covering AI inference and compute orchestration, and that it will receive a non-exclusive, perpetual license to Clarifai's modern AI inference and reasoning stack, to strengthen Nebius Token Factory. Zeiler joins Nebius as SVP, Research, leading work on multimodal agentic reasoning, world models, token efficiency and long-term memory. Commercial terms were not disclosed. Recent Clarifai announcements also include a partnership with Arcee AI to serve as primary inference provider for its Trinity family of US-built open-weight models, and a joint demonstration with Vultr of GPU inference performance at the NVIDIA GTC AI Conference.