If you want a self-hosted development environment that can handle single sign-on and automated shutdown of idle resources to minimize costs, Coder is worth a look. It lets you create secure, high-performance development environments that can be adapted to your infrastructure needs. With features like AI/ML development support, automated resource shutdown, and secure environments, Coder hopes to help you lower cloud costs and shorten build and test times. It can run on a variety of infrastructure, including virtual machines, Kubernetes pods and Docker containers, and is available as open-source software and with enterprise pricing starting at $5 per user per year.
Another interesting project is Auth0, which offers powerful authentication and authorization abilities, including Single Sign On and Multifactor Authentication. It can be easily integrated with many programming languages and frameworks, so it's a good tool for managing user identities across many different organizations. Pricing isn't clearly disclosed, but Auth0 offers a free plan so you can try it out and assess its utility before moving to a paid plan.
For running AI workloads, dstack is a good option. It automates infrastructure provisioning and supports a variety of cloud providers, including AWS, GCP, Azure and on-prem servers. dstack offers several deployment options, including self-hosted and managed versions, and can help you cut costs by using lower-priced cloud GPUs. That makes it a good tool for managing and deploying AI models while keeping costs in check.
If security is a big concern, check out SecureAuth, which offers a passwordless continuous authentication platform. It offers continuous authentication, invisible MFA and single sign-on, so it's a good option for protecting both workforce and customer identities. Moving to passwordless authentication can dramatically reduce credential-level attacks and improve overall security.