AlexaDost

• 

  Intial

• 

  questioning it

• 

  analyzing

• 

  it remembers

• 

  add catchup questions

• 

  gives perfect answer

• 

  answer

Inspiration As a student, juggling multiple assignments, upcoming deadlines, and study notes across different platforms is a constant challenge. When exploring the capabilities for the Amazon Developer Hackathon (Alexa+ track), I wanted to build something genuinely practical for everyday student life. That led to AlexaDost—an intelligent, voice-friendly study sidekick that brings smart academic planning and revision directly into the hands of students.

What it does AlexaDost is a custom Model Context Protocol (MCP) server coupled with a browser-based simulator that equips Alexa+ with 9 specialized voice tools. It allows students to:

Manage Deadlines: Add, track, and parse assignments or exam dates using natural speech.

Automate Study Schedules: Run an earliest-deadline-first scheduling algorithm to build optimized weekly plans and catch scheduling conflicts early.

Spaced-Repetition Revision: Add flashcards and review them using the SM-2 algorithm for efficient memory retention.

Daily Briefings: Get an instant voice summary of upcoming tasks, due flashcards, and daily study goals in a single command.

How we built it Backend & Server: Built using Python and FastAPI, following the latest Model Context Protocol (MCP) specification over Streamable HTTP.

Database & Logic: Uses SQLite for reliable local data persistence, integrated with custom algorithms for date parsing, deadline prioritization, and spaced repetition.

Simulator & UI: Features an interactive browser-based simulator so users can test voice interactions, view tool responses, and track real-time MCP server logs.

Challenges we ran into Implementing a robust MCP server spec while handling complex, open-ended natural language time inputs (like "due in 3 days 6 hours") required careful regex and date-parsing logic. Designing the browser simulator to cleanly display both voice outputs and underlying tool-call payloads simultaneously also required fine-tuning the full-stack architecture.

Accomplishments that we're proud of Successfully creating 9 fully functional, voice-optimized tools tailored specifically for student productivity.

Implementing a complete, working MCP server implementation that passes rigorous local tests.

Delivering a clean, end-to-end user experience with an interactive simulator ready for demo.

What we learned Building AlexaDost provided deep hands-on experience with the Model Context Protocol (MCP) ecosystem, standardizing tool definitions, and structuring context windows to ensure voice-activated agents return concise, conversational answers.

Built With

• css
• fastapi
• html
• javascript
• mcp
• python
• sqlite