Generate, validate, and explain Terraform configurations and provider docs via HashiCorp's official Terraform MCP. Useful for infra-as-code agents.
No configuration required
Deploy and start querying — nothing to connect.
Light — starts in seconds
Sessions run up to 60 minutes before the sandbox is reclaimed.
60+ AI models
Chat with 60+ AI models on the same workflow — switch to a different model mid-conversation and re-run the same prompt, or use Compare mode to put several side-by-side.
How models use it and what it is built for.
HashiCorp’s official Terraform MCP server helps a model write correct Terraform — generating configurations, validating syntax and looking up provider documentation at query time.
Provider drift is the problem it solves. Terraform providers change resource arguments constantly, and a model working from training data confidently emits arguments that were renamed or removed two releases ago. Live provider docs mean the generated config matches the provider you are actually running.
It requires no credentials, because it never touches your infrastructure. It reads documentation and reasons about configuration — it does not plan or apply.
Concrete operations exposed as tools.
In order. Each step assumes the previous one worked.
No configuration and no credentials. The server reads public provider documentation, so there is nothing to connect.
Say which provider you are targeting and, ideally, which version. That is what makes the lookup precise rather than generic.
Generated Terraform is a draft. Run plan yourself and read the diff — this server has no visibility into your state file or your existing infrastructure.
Paste any of these into Agent Studio once the server is connected.
Write Terraform for an S3 bucket with versioning, encryption and public access blocked.
What arguments does the aws_ecs_service resource accept in the current AWS provider?
Explain what this Terraform module does and what it will create.
Convert this inline resource block into a reusable module with variables.
Known constraints, stated plainly.
Deploy the server hosted, then watch which tools a model actually reaches for — with full JSON input and output on every call. Switch models mid-conversation to compare how each one uses the same server.
About the Terraform MCP server.
No, and that is deliberate. It generates and validates configuration but never runs plan or apply, so it needs no cloud credentials and cannot change anything. You review the output and apply it yourself.
Yes — it is the official Terraform MCP server, which is why its provider documentation lookups track current releases rather than a scraped snapshot.
Because provider arguments change often and models confidently emit deprecated ones. Live documentation lookup is the difference between config that plans cleanly and config that fails on an unknown argument.
No. It never contacts a cloud provider — it reads documentation and reasons about HCL. That is what makes it safe to run with no configuration at all.