Komodo services all you to easily serve models that scale up and down with traffic. Just define your service, and we’ll take care of the autoscaling and load balancing.

​
Key Features

  • Custom, Reproducable Environments: Define your own custom environment with all the dependencies you need in a simple YAML file
  • Autoscaling: Komodo will scale up and down replicas for your service based on traffic, so that your models are production-ready at any scale.
  • Scale-To-Zero: Enable serverless-style model serving with scale-to-zero capabilities. When your model has no traffic, we’ll scale your replicas down to zero so that you’re not paying for idle compute.
  • Automatic Git Access: Simply add the public SSH key from the settings page to your Github/Gitlab account, and all your services will have access to pull your private repos
  • Multi-Cloud Scaling: When Komodo scales up your replicas, it will provision replicas in any of the cloud accounts you have connected, ensure maximum GPU availability at the lowest prices.