massongit/open_university_japan_syllabus_mcp_server

Open University of Japan (Face-to-Face Course) Syllabus MCP Server

| English

A Model Context Protocol (MCP) server for the Open University of Japan. It provides syllabus data for face-to-face courses. The server automatically retrieves and parses PDF files from the university's official site. Data is then stored in a structured SQLite database. Functions as an MCP tool that can be directly searched and utilized by AI tools like Claude Code.

Key Features

• PDF Data Extraction: Automatically extracts table data from syllabus PDFs. Uses pdfplumber for face-to-face courses on the web
• SQLite Database: Structured storage of extracted syllabus data in a searchable format
• MCP Server: Provides search APIs directly accessible from AI tools. Functions as a Model Context Protocol server
• Multi-faceted Search Functions:
  • Regional and study center-based search
  • Cross-field keyword search (course names, instructors, course overviews, etc.)
  • Academic level-based search (numbering system support)

Requirements

• Python 3.13 or higher
• uv package manager

Setup

# Install dependencies
uv sync

Usage

MCP Server Registration

For Claude Code, register as an MCP server with the following command:

claude mcp add open-university-japan-syllabus -- uvx --from git+https://github.com/massongit/open_university_japan_syllabus_mcp_server open-university-japan-syllabus-mcp-server

Database Generation

You can generate the database with the MCP server's generate_db command. To create the database via CLI, run the following command:

uv run generate_db.py

Note: This command takes approximately 16 minutes to complete. Rate limiting is set to 1 request per minute to avoid overloading the university's servers.

Available MCP Tools

The MCP server provides the following tools:

get_version

Retrieves MCP server version information.

generate_db

Generates the database with the latest syllabus data.

Note: This operation takes approximately 16 minutes due to rate limiting.

get_face_to_face_courses_syllabus

Searches syllabi for face-to-face courses. Flexible searching is possible with the following parameters.

• places: Filter by region/location
  • 7 regional blocks:
    • Hokkaido-Tohoku, Kita-Kanto-Koshinetsu, Minami-Kanto
    • Tokai-Hokuriku, Kinki, Chugoku-Shikoku, Kyushu-Okinawa
  • Specific study center names or satellite space names
• keywords: Keyword search
  • Full-text search across course names, course categories, instructors, and course overviews
• numbering_hundred_digits: Filter by academic level
  • 1: Foundation subjects (100-199)
  • 2: Introductory subjects (200-299)
  • 3: Specialized subjects (300-399)
  • 4: Comprehensive subjects (400-499)

Development

Development Environment Setup

Install development dependencies and pre-commit hooks.

# Install development dependencies
uv sync --group dev

# Install pre-commit hooks (required to install gitleaks)
uvx pre-commit install

Architecture

File Structure

• src/mcp_server.py: FastMCP-based MCP server implementation
• generate_db.py: Database generation entry point
• src/lib.py: Core PDF processing and data transformation library
• src/rate_limiter.py: Token Bucket algorithm implementation for HTTP request rate limiting

Data Processing Flow

1. PDF Retrieval: Automatic download of PDFs from 7 regional blocks. Retrieved from the Open University of Japan official site
2. Rate Limiting: Limited to 1 request per minute using Token Bucket algorithm
3. Table Data Extraction: Table structure analysis and data extraction using pdfplumber
4. Data Normalization: Japanese text normalization and study center information supplementation
5. Database Storage: Structured storage in SQLite database
6. MCP API Provision: Dynamic search query generation and JSON format responses

Technical Specifications

• Target Data: Open University of Japan (Face-to-Face Course) Syllabi
  • Note: Live Web courses are not included.
• Update Frequency: Manual execution
• Search Limits: Maximum 20 responses per query
• Database: SQLite (cross-platform support using platformdirs)