Spacelift announced Spacelift Flows, extending its platform beyond infrastructure deployment to day 2 operations, where platform teams spend the majority of their time running, maintaining and troubleshooting infrastructure already in production.
AI-assisted development is increasing the volume and velocity of infrastructure reaching production, but automation and governance have historically stopped short of this operational phase. Spacelift, known for bringing automation and governance to infrastructure as code (IaC) and GitOps deployment, now applies its automation, policy enforcement and audit control expertise to day 2 operations, helping platform engineers keep pace with AI-generated code without becoming the bottleneck.
Spacelift Flows carries automation past provisioning into operations
Infrastructure teams have spent the past decade automating provisioning. However, the operations that occur after a deployment have remained manual and ad hoc. Making matters worse, platform engineers have had no shared view of what automations are running and no record of execution when something fails. That changes with Spacelift Flows. It extends Spacelift’s automation and governance, as well as its AI-powered Spacelift Intelligence layer, to day 2 operations.
“Platform teams are absorbing more infrastructure requests than ever as AI accelerates how quickly their developers write code,” said Pawel Hytry, CEO of Spacelift. “Our customers already manage their IaC pipelines with Spacelift, and Flows extends that same automation, visibility and governance to everything that happens after. When the work after deployment is governed and automated, features ship faster, fewer things break, and platform engineers spend their days on what’s next instead of firefighting.”
The Spacelift infrastructure orchestration platform
The Spacelift platform now manages the full infrastructure lifecycle on a single control plane, with one governance layer and one AI layer spanning deployment and operations. The IaC and GitOps deployment component is now called Spacelift Deploy, and the new operations component is called Spacelift Flows. Spanning both components is Spacelift Intelligence, an intelligence layer that helps understand, design, govern and even deploy infrastructure with natural language.
Spacelift now gives platform teams a single place to see every infrastructure automation running in their organization, with a step-by-step execution trail, real-time status and the ability to replay a workflow either in whole or in parts.
Governance and audit that don’t stop at provisioning
Every control that applies to infrastructure deployed with Spacelift now also applies to what happens after it. Policy, role-based access control and audit history work the same way on a provisioning run and on an operational workflow. Deployment events can trigger operational workflows and vice versa.
Clear ownership for agentic infrastructure
Platform teams increasingly want agentic workflows, where an agent, not a person, makes the change. Spacelift Flows gives agents a governed path into infrastructure. Teams can expose internal databases, monitoring systems and operational APIs as MCP servers that agents can call, with private network access and approval gates. Each agent authenticates as itself and appears in the audit trail as its own user, rather than borrowing an engineer’s credentials or using a shared key. Every action is attributable to the agent that took it. The workflows themselves execute deterministically, running the same way every time.
The platform team at Discovery Education has deployed Spacelift Flows and is realizing the various benefits of agentic workflows after infrastructure deployment. Toby Norman, partner success manager, said “I hadn’t built an MCP server for my infrastructure before. It went very smoothly with Spacelift Flows. In a few hours I had several tools running and a proof of concept that helps developers debug their services on our platform.” Parker Cameron, senior manager, cloud engineering, said “With Flows, we had something working in a few hours, instead of four weeks later and still struggling with auth, which is usually the pattern.”
