MCP Server: Terraform AWS Provider Docs

A Model Context Protocol (MCP) server built with Deno and TypeScript, designed to provide contextual information related to Terraform AWS Provider.

Table of Contents

• MCP Server: Terraform AWS Provider Docs
  • Table of Contents
  • Overview
    • What is the Model Context Protocol (MCP) and how does it work?
    • Why?
  • Tools
  • Getting Started
    • Install and Use with Claude Desktop
      • Using Deno
      • Using Docker
    • Installing in IDE/Editor(s)o
      • Install on Cursor
      • Install on Windsurf
      • Install on VSCode
  • Developing & Contributing
    • Run it directly from JSR
    • Debugging & Troubleshooting
    • Using Docker
  • Roadmap
  • Contributing
  • Security
  • License

Overview

This server acts as an MCP server, exposing tools and resources that allow AI agents or other MCP clients to query information about Terraform AWS Provider information, such as:

• ✅ Resources documentation.
• ✅ Provider's configuration, including ephemeral resources, guides, and functions.
• ✅ GitHub Issues (opened, closed, and all)
• ✅ AWS Resources examples

What is the Model Context Protocol (MCP) and how does it work?

The Model Context Protocol (MCP) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. Whether you're building an AI-powered IDE, enhancing a chat interface, or creating custom AI workflows, MCP provides a standardized way to connect LLMs with the context they need.

— Model Context Protocol README

---

Why?

When writing IaC, or designing terraform modules, it's often required a very good knowledge, understanding and context in the actual AWS resources, features, and capabilities in order to design a production-grade module, with stable interfaces, composable, and reusable.

This MCP server is designed to provide just that, with the latest documentation, issues, and examples from the Terraform AWS Provider registry site. Always up-to-date, and always from the source.

Tools

Important

All tools require a valid GitHub token set as an environment variable: GITHUB_TOKEN, GH_TOKEN, or GITHUB_PERSONAL_ACCESS_TOKEN.

Currently, the following tools are available (more to come, or feel free to submit an issue or PR):

Tool Name	Purpose	Inputs	Outputs	Use Case
list-resources	List all AWS resource documentation files with metadata	None	Array of resource objects with id, subcategory, page_title, description, resource, resource_description, source, file_path	Discover all AWS resources supported by the Terraform AWS Provider, obtain file names/paths for further queries, build dashboards, summaries, or analytics.
get-resource-doc	Fetch a single AWS resource documentation file by name or fuzzy search	aws_resource (string, optional): Resource name or description file_name (string, optional): Exact file name (takes precedence)	Parsed and formatted documentation for the requested resource, with all metadata fields and full markdown body	Retrieve detailed documentation for a specific AWS resource, resolve ambiguous or natural language queries, provide LLMs or clients with structured resource information.
list-datasources	List all AWS datasource documentation files with metadata	None	Array of datasource objects with id, subcategory, page_title, description, datasource, datasource_description, source, file_path	Discover all AWS datasources supported by the Terraform AWS Provider, obtain file names/paths for further queries, build dashboards, summaries, or analytics.
get-datasource-doc	Fetch a single AWS datasource documentation file by name or fuzzy search	aws_datasource (string, optional): Datasource name or description file_name (string, optional): Exact file name (takes precedence)	Parsed and formatted documentation for the requested datasource, with all metadata fields and full markdown body	Retrieve detailed documentation for a specific AWS datasource, resolve ambiguous or natural language queries, provide LLMs or clients with structured datasource information.
get-open-issues	Retrieve open issues from Terraform AWS Provider GitHub repo	all (boolean, optional): Retrieve all or first 30 issues	Array of issue objects with ID, title, description, source, state, user, labels, creation/update timestamps, comments	Analyze, triage, or report on current open issues. Build dashboards, correlate issues with documentation, understand if a Terraform behavior is expected or related to a known issue.
get-issue	Fetch detailed information for a specific GitHub issue	issueNumber (number, required): Exact GitHub issue number	Detailed issue object with full metadata including body, timestamps, labels, comments	Investigate a specific issue in detail, typically used after get-open-issues to obtain comprehensive information about a single issue of interest.
list-all-releases	Retrieve all releases from the Terraform AWS Provider GitHub repo	None	Array of release objects with ID, tag, name, author, published date, URL, asset count, and summary body	List all available versions/releases, build dashboards, changelogs, or analytics, correlate provider versions with documentation, issues, or upgrade guides.
get-release-by-tag	Fetch detailed information for a specific release by tag	tag (string, required): Release tag (e.g. 'v5.96.0') include_issues (boolean, optional): Also fetch referenced issues	Release object with full metadata and, if include_issues is true, details for all referenced issues in the release notes	Investigate a particular version in depth, including referenced issues. Use after list-all-releases or directly if the tag is known.
get-latest-release	Fetch detailed information for the latest release	include_issues (boolean, optional): Also fetch referenced issues	Latest release object with full metadata and, if include_issues is true, details for all referenced issues in the release notes	Quickly access the most recent version, check for new features or bug fixes, and optionally see referenced issues.

Getting Started

Install and Use with Claude Desktop

To use this Deno-based MCP server with Claude Desktop, add the following to your claude_desktop_config.json:

Using Deno

{
  "mcpServers": {
    "tf_aws_provider_docs": {
      "command": "deno",
      "args": [
        "run",
        "-A",
        "main.ts"
      ],
      "env": {
        "GITHUB_TOKEN": "<YOUR_TOKEN>"
      }
    }
  }
}

Or, the recommended way, with deno directly from JSR

{
  "mcpServers": {
    "tf_aws_provider_docs": {
      "command": "deno",
      "args": [
        "run",
        "-A",
        "jsr:@excoriate/[email protected]"
      ],
      "env": {
        "GITHUB_TOKEN": "<YOUR_TOKEN>"
      }
    }
  }
}

Using Docker

{
  "mcpServers": {
    "tf_aws_provider_docs": {
      "command": "docker",
      "args": [
        "run",
        "-e",
        "GITHUB_TOKEN=<YOUR_TOKEN>",
        "mcp-terraform-aws-provider-docs"
      ],
      "env": {
        "GITHUB_TOKEN": "<YOUR_TOKEN>"
      }
    }
  }
}

Installing in IDE/Editor(s)o

Install on Cursor

Go to: Settings -> Cursor Settings -> MCP -> Add new global MCP server

Pasting the following configuration into your Cursor ~/.cursor/mcp.json file is the recommended approach. See Cursor MCP docs for more info.

{
  "mcpServers": {
    "tf_aws_provider_docs": {
      "command": "deno",
      "args": ["-A", "jsr:@excoriate/mcp-terraform-aws-provider-docs@latest"]
    }
  }
}

Install on Windsurf

Add this to your Windsurf MCP config file. See Windsurf MCP docs for more info.

{
  "mcpServers": {
    "tf_aws_provider_docs": {
      "command": "deno",
      "args": ["-A", "jsr:@excoriate/mcp-terraform-aws-provider-docs@latest"]
    }
  }
}

Install on VSCode

Add this to your VSCode MCP config file. See VSCode MCP docs for more info.

{
  "servers": {
    "Context7": {
      "type": "stdio",
      "command": "deno",
      "args": ["-A", "jsr:@excoriate/mcp-terraform-aws-provider-docs@latest"]
    }
  }
}

Developing & Contributing

For more details about debugging, testing, and contributing to this project, see DEVELOPER_GUIDE, and CONTRIBUTING.

Run it directly from JSR

You can use the MCP server directly from JSR (Javascript Registry ❤️)

# export your github token
export GITHUB_TOKEN=ghp_xxx...

# run it
deno run -A jsr:@excoriate/mcp-terraform-aws-provider-docs@latest

Debugging & Troubleshooting

if you want to debug it, use the built-in debugger (inspector). There's a justfile recipe to help you out.

# start the mcp server, and the inspector
just inspect

Using Docker

Build the Docker image

docker build -t mcp-terraform-aws-provider-docs .

Run the MCP server in Docker

docker run -it --rm \
  -e GITHUB_TOKEN=ghp_xxx... \
  mcp-terraform-aws-provider-docs

Tip

Replace ghp_xxx... with your GitHub Personal Access Token with appropriate permissions.

You can also use GH_TOKEN or GITHUB_PERSONAL_ACCESS_TOKEN as the environment variable name.

If you want to use a local .env file, you can pass it with --env-file .env.

Roadmap

• Add tool to retrieve provider's configuration.
• Integrated with dedicated backend, to index, and provide advance search/code assistant documentation avoiding GitHub API rate limits.

Contributing

See CONTRIBUTING.md for detailed contribution guidelines, including setup, code style, PR process, and codebase structure reference.

Security

See SECURITY.md for the project's security policy, including how to report vulnerabilities and responsible disclosure guidelines.

License

This project is licensed under the MIT License.