Serify banner

Serify

Created by: Sam Wu
1 devlog
10h 7m 19s

Serify is a context-aware learning reflection engine that analyzes content you've consumed — YouTube videos, articles, PDFs, or personal notes — and generates intelligent, scenario-based questions designed not just to test you, but to diagnose how…

Serify is a context-aware learning reflection engine that analyzes content you’ve consumed — YouTube videos, articles, PDFs, or personal notes — and generates intelligent, scenario-based questions designed not just to test you, but to diagnose how well you actually understand the material.

This project uses AI

Copilot for code completion, learning, and complex tasks.

Repository

README

Loading README...
Sam Wu
Sam Wu worked on Serify
23 days ago
10h 7m logged

2/22/26 Devlog

Project Vision

I really wanted to build something to help me learn faster and excel with less effort in school.
So I started building serify: A context-aware learning reflection engine designed to move beyond simple testing. It uses AI to analyze conceptual depth, identify misconceptions, and map knowledge gaps in real-time.

Core Features Implemented

1. AI-Powered Diagnostic Engine

  • Content Processing: Extracts hierarchical concept maps from raw text, YouTube URLs, or PDFs using Gemini 2.5 Pro.
  • Dynamic Question Generation: Creates retrieval, application, and misconception-probe questions tailored to the extracted concepts.
  • Real-time Answer Analysis: Analyzes student responses for factual accuracy and conceptual depth without blocking the user flow.
  • Feedback Synthesis: Generates a comprehensive report including a summary of grasp, strength maps, and actionable focus suggestions.

2. Data & Persistence

  • Session History: Persistent tracking of all historical and active sessions in localStorage.
  • Knowledge Gaps: A centralized summary of identified conceptual gaps rendered on the main dashboard.
  • Storage Utility: Centralized lib/storage.ts for managing the lifecycle of learning sessions.

Technical Architecture

  • Framework: Next.js (Pages Router)
  • AI Backend: Google Gemini API (using 2.5 pro) via @google/generative-ai
  • Styling: Vanilla CSS with modern CSS variables for theming and Glassmorphism effects.
  • Icons: Lucide React
  • Authentication: Firebase/Custom integration (ready for bridge to Supabase).
Attachment
Attachment
Attachment
0