Hangul

Hangul-MCP is a Korean language utilities service built on the MCP (Model Chaining Protocol). It exposes a set of tools for processing Korean text, including spelling/grammar checking, character and word counting, resume-specific review, and romanization according to the National Institute of Korean Language romanization rules. The server integrates with client applications via MCP messages, allowing you to request a specific tool (for example, check_korean_grammar or romanize_korean) and receive structured responses with results and suggestions. This makes it suitable for embedding within larger text-processing pipelines, chatbots, or editorial workflows where Korean language quality is important.

To use the tools, configure your MCP client to connect to the Hangul-MCP server and invoke the desired tool by name. Each tool corresponds to a distinct function: count_korean_letters returns character and word counts for Korean text; check_korean_grammar analyzes grammar and spelling and returns suggested corrections; resume_review combines length checks with spelling feedback tailored to resumes; romanize_korean converts Korean text to romanized form following the standard romanization rules. You can batch requests or perform single-shot analyses depending on your integration pattern.

Prerequisites:

• Docker installed and running (recommended for this deployment).
• Access to a terminal or CI environment to pull and run the container.

Installation steps:

1. Pull the Hangul-MCP Docker image (or build locally if you have a custom image):

   docker pull alfex4936/hangul-mcp

2. Run the MCP server container:

   docker run -i alfex4936/hangul-mcp

3. If you prefer to run locally with the exact binary (not containerized), clone the repository and build the Go binary as described in the README:

   git clone https://github.com/Alfex4936/Hangul-MCP cd Hangul-MCP go mod download go build -o kogrammar

4. Start the binary directly (example, adjust path as needed):

   ./kogrammar

5. Ensure your MCP client is configured to connect to the running server (host and port as applicable in your environment).

Tips and common issues:

• If you’re using the Docker approach, ensure your environment allows container networking and that the container has access to any required resources or models used by kospell underneath.
• For production deployments, consider exposing metrics and health endpoints, and configure a stable logging approach (LOG_LEVEL can help).
• The server relies on the kospell library for spelling checks; ensure compatibility with the Korean language data you intend to process.
• When wiring into MCP, verify the tool names exactly match (count_korean_letters, check_korean_grammar, resume_review, romanize_korean) and structure your requests to receive consistent JSON responses.
• If you encounter encoding issues, ensure your environment uses UTF-8 encoding for all inputs and outputs.