New guide covers provisioning a governed Databricks workspace baseline with Pulumi, including cluster policies, notebooks, secret scopes, permissions, and automated jobs.
Pulumi
pulumi.comInfrastructureTracking since Oct 2024
$
npx @buildinternet/releases get pulumiAd-hoc resource snippets in state; Output.recover for Node.js
↗This release4 featuresNew capabilities2 enhancementsImprovements to existing features4 fixesBug fixesAI-tallied from the release notesPulumi · v3.248.0
Ad-hoc resources can now be tracked as PCL snippets in state, and Node.js SDK adds Output.recover to catch exceptions in outputs. Fixed a panic when binding a schema that references an uninstalled plugin, and generated programs now guard conditionally-created resources.
Personal AI agents had their breakout this year. [OpenClaw](https://www.p...
New blog post walks through building a standardized Cloudflare edge baseline with Pulumi, covering DNS records, WAF custom rules, a Worker canary, and Zero Trust Access policies.
Neo code reviews analyze IaC against live infrastructure
↗This release1 featureNew capabilitiesAI-tallied from the release notesPulumi introduced Neo code reviews, now in public preview, which analyze pull request changes against live cloud infrastructure and preview output. Reviews post as comments on GitHub PRs, respect existing RBAC and guardrails, and are free during preview before GA on July 1, 2026.
OAuth auto-refresh on 401; registered resources retrievable from mock monitor
↗This release10 featuresNew capabilities9 enhancementsImprovements to existing features6 fixesBug fixesAI-tallied from the release notesPulumi · v3.247.0
CLI now auto-refreshes OAuth tokens on 401 instead of returning "login required" errors. Registered resources can be retrieved from the mock monitor for test assertions. Also adds --provider flag to pulumi do, --raw flag to pulumi config set, and Output.recover to Python SDK.
Workshop writeup covers six recommendations for Kubernetes on Azure, including enabling Cilium on AKS, picking a team-aligned language, using AI-assisted debugging to route around Docker Hub rate limits, splitting infra from workload into stacks, and managing state with Pulumi Cloud.
Log listing and removal added; `--stateless` flag for `pulumi do`
↗This release5 featuresNew capabilities4 enhancementsImprovements to existing features9 fixesBug fixesAI-tallied from the release notesPulumi · v3.246.0
Added pulumi logs ls and pulumi logs rm commands to manage automatic log files. A breaking change deletes all backend config upon logout. The pulumi do command now requires a --stateless flag for direct-provider behavior. Fixed backend compatibility issues with gocloud.dev and schema-declared output fields in PCL programs.
Article explores loop engineering, a shift from manually prompting coding agents turn-by-turn to designing automated systems that prompt agents continuously. Covers five building blocks (automations, worktrees, skills, connectors, sub-agents) plus memory infrastructure, and discusses verification, token costs, and knowledge maintenance as ongoing responsibilities.
Pulumi Deployments can now trigger pulumi up automatically when a git tag is pushed, with glob-based tag filters to control which tags deploy. The feature is available across GitHub, GitLab, Bitbucket, Azure DevOps, and custom VCS integrations, and sets the PULUMI_CI_TAG_NAME environment variable for use in pre-run commands or Pulumi programs.
Stack init renamed to new; onError hooks no longer dropped
↗This release4 featuresNew capabilities1 enhancementImprovements to existing features3 fixesBug fixesAI-tallied from the release notesPulumi · v3.245.0
The pulumi stack init command is now named new. Fixed mergeOptions dropping onError hooks from ResourceOptions in the Node.js SDK, and the HCL language runtime is now downloaded on demand instead of bundled. Plugin installation now displays download and unpack progress, and pulumi neo supports arrow keys to scroll through prompt history.
New guide shows how to run Google's open-weight Gemma 4 12 B model locally on a Mac using llama.cpp for native inference, k3d for local Kubernetes, Pulumi for infrastructure as code, and Tailscale for secure access. The setup achieves about 20 output tokens per second on an M3 Max MacBook Pro with 36 GB RAM and supports multimodal input including images and audio.
New guide shows how to build an environment factory on Amazon EKS using Pulumi and vCluster, adapting Deloitte's architecture that achieved 89% faster testing environment provisioning. The pattern consolidates multiple isolated, ephemeral Kubernetes environments into a single host cluster using EKS Auto Mode and virtual clusters, reducing provisioning time and operational overhead while maintaining strong tenant isolation.
New guide covers running multiple AI coding agents in parallel on infrastructure using Pulumi components, CrossGuard policies, review stacks, and ESC environments for isolation. The playbook adapts five patterns from application-code parallelism—issue as spec, plan/build/validate loop, parallel worktrees, fresh-session review, and self-healing layer—to infrastructure work, with stack-scoped isolation replacing port and database isolation.
Snapshot integrity fixed with up --refresh; provider errors now forwarded to hooks
↗This release6 fixesBug fixesAI-tallied from the release notesPulumi · v3.244.0
Fixed a snapshot integrity issue that occurred with up --refresh and ensured all provider errors are forwarded to error hooks for retry. Also fixed top-level flags like --logtostderr being recognized when using pulumi do, distinguished multiple packages with the same plugin, improved trace spans for cancel RPCs during shutdown, and fixed requirements.txt merging into pyproject.toml when using uv.
Service Provider now auto-generates from OpenAPI; RBAC and audit logs as code
↗This release3 featuresNew capabilitiesAI-tallied from the release notesThe Pulumi Service Provider v1.0 is now generated directly from the Pulumi Cloud OpenAPI specification, enabling new resources to ship in the provider the same release they reach Pulumi Cloud. v1.0 adds fine-grained RBAC as code (custom roles, organization membership, team role assignments), Pulumi IDP as code, and audit log export as infrastructure as code.
A detailed guide on wiring context and tooling around Claude Code to improve agent performance in real codebases. Covers seven harness layers: lean CLAUDE.md files, hooks for self-improvement, progressive-disclosure skills, plugin distribution, LSP navigation, MCP server extension, and subagents for splitting work. The post emphasizes that the harness—not the model version—determines agent performance.
New guide surveys AI infrastructure tools across two categories: compute and MLOps platforms (CoreWeave, Modal, Weights & Biases, MLflow, hyperscaler offerings) and AI-powered infrastructure management (Pulumi Neo, Firefly, env0, Spacelift, code assistants). Distinguishes infrastructure for AI from AI-powered infrastructure management and covers vendor strengths, use cases, and trade-offs.
pulumi do gains resource support; neo UX improved
↗This release8 featuresNew capabilities1 enhancementImprovements to existing features4 fixesBug fixesAI-tallied from the release notesPulumi · v3.243.0
Resource support is now available in the pulumi do command. The pulumi neo agent UI improved with Esc key handling to clear input or cancel turns. Also aliased pulumi new to pulumi project new, added a List method to the Go plugin.Provider interface, and made pulumi project new -y write minimal project files without templates.
pulumi do command for ad-hoc cloud resource operations
↗This release1 featureNew capabilitiesAI-tallied from the release notesPulumi introduces pulumi do, a new command for direct resource operations that create, read, update, delete, and query cloud resources from the terminal without requiring a Pulumi project, code, or state. The command supports thousands of Pulumi-backed providers and works with provider functions to look up existing resources, collapsing multiple setup steps into a single operation.