Terraform

HashiCorp's official Terraform MCP bridges Claude to the Terraform Registry, letting you query module docs, pull provider schemas, and validate configs without leaving your editor. It's a Docker-based server that runs locally and surfaces the same data you'd otherwise tab over to the registry website to find. I've used it most when scaffolding new infrastructure. You can ask Claude to look up the latest AWS provider arguments, pull example snippets from popular modules, or check what outputs a community module exposes before you wire it into your stack. The workflow is: describe what you're building, let Claude query the registry through the MCP, then paste the generated HCL into your .tf files. It's faster than hunting through registry docs manually, and you get context-aware suggestions based on what you've already written. The Docker requirement is both a strength and a friction point. You're not installing npm packages or Python deps, which keeps your system clean, but you do need Docker running. The server itself is stateless and quick to spin up. No API keys, no auth, just the public registry data piped through. Quirks: it won't execute Terraform commands or manage state. This is purely a documentation and discovery tool. If you want to run `terraform apply` from Claude, you'll need a different MCP or a shell tool. Also, it's read-only against the registry, so you can't publish modules or manage your own registry entries through it. Skip this if you're not already writing Terraform daily. It's for people who live in .tf files and want to reduce the number of browser tabs open while they work. If you're just learning Terraform or only touch it occasionally, the overhead of running a Docker server isn't worth it. But if you're a platform engineer or SRE who writes infrastructure code every day, this is a genuine time-saver. The official backing from HashiCorp means it'll stay current with registry changes, which matters more than you'd think.