Stardrift

Login page
Dashboard page
Upload
Learning Mode
Asking Mode

Inspiration

For students with ADHD, a standard 50-page PDF isn't just a document—it’s a "Wall of Awful." This phenomenon, known as task paralysis, causes severe emotional distress and burnout. We wanted to build a "Light Side" solution that transforms terrifying documents into manageable, interactive Galactic Sprints, reducing the cognitive load and anxiety associated with traditional studying.

What it does

Stardrift is an AI-powered study companion that uses a Star Wars-themed interface to gamify learning:

PDF to Holocron: Uses Gemini 1.5 Flash to deconstruct complex PDFs and messy handwriting into "Bite-sized Sprints".

Voice-First Mentorship: Converts study material into high-fidelity audio using ElevenLabs (the "George" mentor voice) to provide a calming, hands-free learning experience.

Socratic Dialogue: Features a "Jedi Master" chat tab that uses the Socratic method, asking the student questions and providing supportive feedback based on their voice or text answers.

Our Features Include:

Transform lecture slides, textbooks, and handwritten notes from pdf form to detailed, explanatory audio lessons
Engage in interactive voice conversations with a “Jedi Master” tutor to deepen understanding
Generate quizzes based on uploaded slides
View full lecture transcripts
Download AI-generated notes
Switch between multiple audio languages
Upload and convert handwritten notes

How we built it

The Brain: Google Gemini 1.5 Flash handles the multimodal analysis of documents and powers the interactive tutor logic.

The Voice: ElevenLabs API provides low-latency, human-like narration to make the content accessible and engaging.

The UI: A React-based frontend styled with a "Star Wars" aesthetic, featuring Holocrons instead of folders to create an immersive, low-stress environment.

Challenges we overcame

A major hurdle was navigating API permissions and "unusual activity" flags on the free tier. We optimized our request architecture to ensure the Scribe v2 (Speech-to-Text) and TTS features worked seamlessly within the hackathon's constraints.

Accomplishments that we're proud of

We successfully integrated a complex multimodal pipeline that takes raw, "scary" data and turns it into a supportive, vocal conversation. Seeing a student go from "I can't start this" to "I'm on a mission" is our biggest win.