The open source stack for
machine learning engineering

Cortex helps you deploy, manage, and scale machine learning in production

Any model

Deploy TensorFlow, PyTorch, and other models as web APIs with one command.

Model serving

At scale

APIs automatically scale to handle production workloads.

Infrastructure automation

Any team

Cortex is open source and available for you to use today.

Applications built on Cortex

A production platform, designed for machine learning engineering

Python Predictor API

Define prediction APIs in Python

Cortex allows you to define prediction APIs for any model—regardless of framework—using an intuitive Python interface.

Cortex YAML configuration

Configure deployments with YAML

Cortex uses declarative YAML to configure deployments. Stop worrying about Kubernetes, Docker, and model servers. Let Cortex do that for you.

Cortex CLI for deploying and managing APIs

Deploy APIs with a simple CLI

Cortex's CLI provides intuitive commands for deploying and monitoring APIs locally and on AWS.

Applying machine learning engineering to real world problems


Treating COVID

PostEra uses machine translation to help manufacture new molecules for drugs, including proposed treatments for COVID.

Check out PostEra

Changing commerce uses computer vision and NLP to create an API capable of automatically categorizing and tagging millions of products according to a hierarchical taxonomy.

Check out

Reimagining sports

Stats Perform uses machine learning to build predictive analytics products for sports organizations, including the NBA, MLB, NFL, La Liga, and more.

Check out Stats Perform

Run Cortex yourself, or let us manage it for you

open source

Run locally or host on your AWS account


Spot instances

Rolling updates

Log streaming

Prediction monitoring

Get started


Request early access to fully managed Cortex, including enterprise features and support.

Contact us