Hive Models Guide

Welcome to Hive! You will soon be on your way to harness the power of Hive’s AI-as-a-service platform.

This guide shows how to integrate Hive’s cutting-edge machine learning models — known as Hive Models — into your software platform to extract groundbreaking insights.

For a full overview of Hive's products and services, please visit the

Overview of Hive Models

Companies often spend years and millions of dollars building machine learning models for specific problems, only to realize unsatisfactory performance. At Hive, our mission is to power innovators to solve problems with practical AI solutions and data labeling, grounded in the world's highest quality visual and audio metadata.

Hive is comprised of two divisions, Data and Models. Hive Data leverages one of the largest distributed workforces in the world, with over two million users making millions of daily judgements to label data. The proprietary training data labeled by the workforce enables Hive Models to develop models that outperform competitors such as Amazon Rekognition and Google Cloud Vision.


Hive offers a set of trained models that are readily available to interface, as well as custom model development services to address customers' unique use cases. For an overview of Hive's model offerings, please visit the Hive Models overview page.

Main types of models