Anyscale

Anyscale belongs in the catalog because it solves a very real production problem for teams building with Ray and large-scale AI workloads. The checked official site positions it as a platform for running and scaling ML and AI workloads across clouds and on-prem infrastructure, with a particularly strong connection to Ray for distributed compute, training, batch inference, and online serving. That makes it relevant to serious engineering teams that have moved beyond local experimentation and need a platform story for operational AI systems.

It also deserves inclusion because the public pricing posture is concrete enough to be buyer-useful without pretending the platform is a simple seat-based SaaS product. The checked pricing page explicitly says buyers can start with pay-as-you-go billing, no credit card is required to get started, additional compute beyond a $100 credit requires a payment method, and invoicing or usage commits are available for larger customers. The page also explains that billing is based on Anyscale Credits and points buyers toward marketplace purchasing on AWS, GCP, and Azure.