Lexicon AI — Academic Vocabulary Journal

Lexicon AI is an intelligence-augmented repository designed for scholars, students, and lifelong learners. Built with a high-contrast, SAT-inspired aesthetic, it transforms the exhausting process of vocabulary acquisition into a structured, AI-powered workflow.

🏛️ The Genesis: A Personal Struggle

Lexicon AI was born from the frustrating reality of SAT preparation.

Standard flashcards and static word lists often fail because they lack situational context. I found myself memorizing definitions for words like recondite or garrulous, only to forget them minutes later because I couldn't "see" them in a real sentence. This project is my solution: an automated system that doesn't just store words, but builds a living, semantic world around them.

✨ Core Features & Evolution 2.0

🏛️ 1. Scholarly Archives (Pre-built Collections)

Access curated, high-stakes vocabulary sets designed for immediate mastery.

IELTS Levels: Progressive collections from A1 to C2.
SAT Mastery: Targeted datasets like the College Panda 400 and Erica Vocabulary lists.
Token-Based Unlocking: Spend your hard-earned tokens to permanently unlock advanced academic archives.

📄 2. AI Document Extractor (Harvesting Engine)

Transform your reading materials into structured journals instantly.

Multi-Format Support: Upload PDF, DocX, or TXT files.
Gemini-Powered Extraction: Lexicon uses Gemini Pro to harvest the most academically significant words from your documents.
Seamless Integration: Review and edit harvested words before they are formalized into your permanent repository.

🤖 3. Intelligent AI Assembly

Input a raw list of words, and the Assembly Engine (gemini-3-flash) handles the rest:

Simplified Definitions: Explaining the concept, not just the word.
Semantic Equivalents: Grouping synonyms for better recall.
Memorable Context: Generates unique, episodic memory triggers.

📖 4. Situational Synthesis (Context Learning)

Using AI, Lexicon can synthesize an entire collection into a single, coherent narrative.

Narrative Mastery: Weaves your vocabulary into a story or scholarly article.
Visual Highlighting: In "Context Mode," words feature tooltips showing how lexemes interact.

🔊 5. AI-Synthesized Pronunciation

Leveraging high-fidelity TTS, the system provides audio for every word to bridge the gap between reading and speaking.

🔍 6. Global Lexical Search

The Unified Repository Search indexes every word across every journal you have ever created.

🛠️ System Reliability & UX Refinements

🪙 Robust Token Ecosystem

Atomic Persistence: Tokens are managed via atomic database transactions, ensuring your balance is never lost or overwritten.
Earn Through Learning: Gain tokens through games, streaks, and sessions.
Motivating Streaks: Earn +10 tokens daily for consistent learning.

🔐 Refined Authentication

Identity Integrity: Pre-flight checks prevent duplicate usernames and provide immediate feedback for existing email accounts.
Silent Failure Prevention: Enhanced UI feedback ensures you never wait for a confirmation email that wasn't sent.

📈 Mastery Architecture

Persistent Progress: Progress is saved even for system collections, allowing you to track your mastery across the entire academic archive.
Visual Indicators: Sophisticated progress bars and circular mastery charts visualize your proficiency.

🏛️ Technology Stack

AI Models
- gemini-2.5-flash: For assembly, extraction, and narrative creation.
- gemini-2.5-flash-tts: For scholarly audio generation.
Backend: Supabase (PostgreSQL) for cross-device synchronization.
Frontend: React 19 + Vite.
Styling: Vanilla CSS for a premium, custom academic feel.

🚀 Deployment & Environment

To deploy this project (e.g., to Vercel), ensure you set the following environment variables in your dashboard:

GEMINI_API_KEY: Your Google Gemini API key for AI functionality
VITE_SUPABASE_URL: Your Supabase project URL
VITE_SUPABASE_ANON_KEY: Your Supabase anonymous key

Local testing

Copy .env.example to .env.local in the project root.
Set the following variables in .env.local:
- GEMINI_API_KEY=...
- VITE_SUPABASE_URL=...
- VITE_SUPABASE_ANON_KEY=...
Start the dev server with npm run dev.

Note: Supabase is configured via the Vite /supabase proxy (see vite.config.ts). If your Supabase tables/policies are not ready, create them using supabase_schema.sql.