Pulumi Context API Unifies Cloud and IaC Topology into Queryable Graph for AI Agents

Pulumi has launched the public preview of the Pulumi Context API, a graph-based interface that connects managed IaC state, cross-stack dependencies, and unmanaged cloud assets discovered outside of IaC into a unified query model. The API processes structured JSON queries comprising anchor, traverse, and return clauses via REST or the Pulumi CLI using the GraphQuery command, available in CLI version 3.243.0 and later. Purpose-built for AI agents, the Context API provides a Markdown schema primer that autonomous agents—such as Pulumi Neo, Cursor, and Claude Code—can retrieve dynamically to understand the query grammar and self-correct on validation errors. The preview is currently available to organizations on Pulumi Enterprise and Business Critical tiers. As cloud environments grow in complexity across multi-stack and multi-account architectures, answering essential operational questions—such as what breaks during a provider upgrade or which stacks consume a specific output—has required manual correlation across disparate state files and cloud consoles. Traditional IaC tools isolate state per stack or workspace, creating visibility silos. By consolidating IaC topology and discovered unmanaged infrastructure into an interconnected graph, the Context API provides deterministic context directly to human operators and autonomous systems, eliminating the guesswork that often leads to breaking changes or orphaned resources. This release aligns with the broader paradigm shift across DevOps toward agentic infrastructure automation and platform-level context engines. As coding agents transition from generating boilerplate templates to executing autonomous changes, they require real-time, governed knowledge of live production topology rather than static codebase views. Pulumi's initiative to expand this graph to encompass Pulumi ESC secrets, RBAC roles, and internal developer platform service catalogs highlights an industry trend: treating infrastructure metadata and dependencies as a primary, queryable control plane. In practice, DevOps and platform engineering teams on supported tiers can leverage the Context API to automate change impact analysis in CI/CD workflows and identify untracked cloud sprawl. Teams embedding AI tools into their infrastructure routines should configure agents to query the schema primer endpoint to enable accurate self-onboarding and syntax correction. Practitioners should note that the API is in preview and the contract may evolve, while access permissions currently mirror existing Resource Search entitlements in the Pulumi console.