NxtFlows Academy · Demo Day

Computer Engineering
Course Planner

An MCP server that gives Claude direct, validated access to a student's real course data — prerequisites, credit limits, and remaining requirements — so academic planning advice is grounded instead of generic.

Jeneen Mohsen Sattouf · Birzeit University, Computer Engineering
Building-an-MCP-for-an-AI-Engine · v1.0.0
View source on GitHub
Why build this

Generic advice vs. grounded answers

Planning a semester means juggling completed courses, prerequisites, credit limits, categories, labs, and projects all at once. MCP is the structured bridge between Claude and this project's real data.

Without MCP

Student → manually explains courses → Claude gives generic advice

  • Completed courses
  • Prerequisites
  • Credit limits
  • Course categories
  • Labs
  • Projects

With MCP

Student → Claude → MCP tools → validated data → personalized, accurate result

  • Completed courses
  • Prerequisites
  • Credit limits
  • Course categories
  • Labs
  • Projects
Demo Day deck

Project slides

The full presentation — why the project exists, how it works, the seven MCP tools, and the live demo prompts.

Slide
Seven tools, one planner

What the MCP server exposes

All tool inputs are validated with Zod before execution — model-supplied input is treated as untrusted.

search_courses
Search the course catalog by query or category.
check_prerequisites
Check eligibility for a given course code.
generate_study_plan
Generate a semester plan under credit & workload constraints.
list_completed_courses
View a student's completed courses.
list_remaining_courses
View a student's remaining requirements.
add_course_to_plan
Add a course to the current plan.
complete_course
Mark a course as completed for a student.