02. THE PROBLEM
Reading comprehension needed an upgrade
Readers often highlight passages or take notes, but these notes get lost and rarely lead to deeper understanding. Most reading apps track progress, and most AI tools summarize content for the user instead of supporting personal insight.
This project aims to help readers remember more of what they read by creating a tool that guides reflection, connects ideas across chapters, and builds a clearer long term memory of books.
03. RESEARCH
Key Insights:
Memory improves when readers explain ideas in their own words.
Emotional relevance, repetition, and structured reflection strengthen retention.
People understand more when they connect new ideas with what they already know.
Metacognition
The act of thinking about one’s own thinking, significantly improves comprehension. When readers pause to evaluate what they understand, identify gaps, and create insights, they retain more. This principle supports the core experience of the app.
Competitive analysis
I studied leading tools across reading and note taking:
Notion: flexible note structures and tagging.
Obsidian: relationship mapping and knowledge graph features.
Bookmory: focus on reading session tracking.
Storygraph: strong in mood tracking and visualization.
Goodreads: social interaction but limited depth for reflection.
04. CONCEPT
A reading assistant that thinks with you, not for you.
Capture reader thoughts while reading.
Highlight relationships between themes, characters, and ideas.
Generate summaries shaped by the reader’s notes, not generic AI text.
Track story elements through a structured data model.
Support guided reflection after each chapter.
Why two modes?
Readers think differently while reading and after reading. Separating these moments creates space for both raw expression and structured understanding. This distinction is intentional and foundational to the experience.
Designed for uninterrupted thinking
Journal Mode keeps the reader’s flow intact. No prompts, no corrections, no interruptions. The system listens quietly so readers can capture thoughts without breaking their connection to the story.
Guided reflection when it matters
Once the chapter is over, Reflection Mode helps readers make sense of what they wrote. Targeted questions reveal patterns, clarify ideas, and strengthen long term memory through structured reasoning.
05. DESIGN APPROACH
06. DESIGNS
All the designs for this prototype were made using Figma, going from a low fidelity idea to a high-fidelity prototype with its own mini design system after multiple rounds of feedback and iterations. The following screens will show the key interactions and visual elements
Home: A quick snapshot of your reading activity
The home screen gives readers a quick snapshot of their current book, recent notes, and weekly reading stats. The layout uses a lot of white space, soft neutral tones, and clear hierarchy to support long form reading and reduce cognitive load.
This screen also pulls data from the structured JSON model built for the project, displaying recent notes, character mentions, and session history. The goal is to offer a lightweight overview without overwhelming the user with options.
Journal Mode: Write freely while you read
This screen lets readers capture thoughts without breaking their flow. The interface is minimal on purpose: soft colors, clear text hierarchy, and no distractions. These notes later become part of summaries and reflection prompts.
The AI takes a listener role in this screen, it never interrupts the user messages and silently takes note of the input to generate summaries and extract details of the chapter.
Session Summary: Your notes organized automatically
After a reading session, the system organizes your notes into a simple structure like summary, characters, and themes. Nothing here is AI invented, all content comes from what the reader wrote. The layout uses clear sections and neutral backgrounds to keep things easy to scan.
07. PROTOTYPE DEMO
Here is a short demonstration of the prototype. It shows the app built with React, using Supabase for the database and OpenAI's API to power conversation prompts, summaries, and the synthesis features. The video walks through how notes are captured, transformed, and used to guide reflection.
This prototype demonstrates the early behavior of the system, including conversation driven reflection, summary generation, and entity tracking across chapters. It also shows how user notes evolve into structured insights.
08. LEARNINGS SO FAR
A reorganized site structure built around real tasks.
A clear and predictable navigation model.
A simplified mega‑menu aligned with user journeys.
Reusable CMS components that reduced inconsistencies.
A unified design system that all departments could follow.
09. NEXT STEPS
User testing with full chapter sequences.
Improved onboarding to explain reflection cycles.
Expanded memory insights across multiple books.
Refined UI hierarchy and long form layout.
Accessibility improvements with semantic labels.