AI Study Assistant — Hackathon Prototype

A beginner-friendly study-planning web app with a local Streamable HTTP MCP server.

Features

• Generate a simple study plan based on subject, level, and available minutes.

• Generate beginner practice prompts.

• MCP tools: create_study_plan and generate_practice_questions.

• Runs locally without an API key.

Related MCP server: MoeReview

Important description

This is a rule-based prototype, not an LLM-powered chatbot. The web interface calls local Flask endpoints; the MCP endpoint exposes tools for an MCP-compatible client. Do not claim it is connected to a live Alexa+ service. For the hackathon, verify the track requirements and clearly describe the demo as a simulated experience if you choose that route.

Requirements

• Python 3.10+

• Internet connection for first-time package installation

Windows setup (PowerShell)

Open a terminal in this folder.

py -m venv .venv
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
.\.venv\Scripts\python.exe server.py

If py is unavailable, use python instead. You do not need to activate the virtual environment. If your current PowerShell location is wrong, first cd into the folder containing server.py.

Open http://127.0.0.1:5000 in your browser. The MCP endpoint is http://127.0.0.1:8000/mcp. Keep the terminal open while using the app; press Ctrl+C to stop.

Troubleshooting

• No such file: make sure the terminal is in the folder that contains server.py. Run Get-Location and dir.

• No module named flask or mcp: run the pip install command above from the same project folder.

• Port in use: close the other process using port 5000 or 8000, then restart.

• PowerShell blocks activation: no activation is required; use .venv\Scripts\python.exe directly.

Submission checklist

• Add a suitable open-source license (MIT is included in LICENSE).

• Push the project to a public GitHub repository.

• Test the app and MCP tools in an MCP-compatible client.

• Record a demo video under 3 minutes.

• Submit the repository URL, video, and required feedback on the hackathon page.