Superwise is model and platform-agnostic – it works with any type of stack. The only thing you need to do to start with model observability is log inferences. Check out our quickstart guides to learn more.

Your first 3 models are entirely free, with up to 1 million predictions per month per model. The community plan has no time, feature, or environmental limits. Use it in production and get immediate visibility into how your models are behaving in the real world. If you need a scale or enterprise plan, contact us at info@superwise.ai.

On scale and enterprise plans, you’ll select a tier depending on the number of models that you need to monitor a month. There are no hidden limitations in any of the plans – all features are fully available to all users in all the plans.

Production predictions can be logged in stream, via batch collectors, local files, or through the UI with CSV files. Check out our documentation to learn more.

You can log actuals together or separately from predictions regardless of the feedback delay. The same unified logging API is used, therefore log in stream or batch similarly to logging predictions data.

A standard Superwise account is self-hosted in the Superwise GCP cloud. In addition, Superwise can be deployed on managed or private cloud instances (AWS, GCP, or Azure) or on-prem. We’re happy to walk you through this – just contact us at info@superwise.ai.

Superwise is SOC2 type2 compliant to request the full report, contact us at legal@superwise.ai.

Superwise is an API-first platform. Everything is accessible via our API to allow organizations not only to export observability data into bespoke BI systems but to fully automate workflows with the API.