Flagship Accessibility Engineering

Lexara Air

Designed for minds that think differently.

AI-powered accessibility system for dyslexic and cognitively diverse learners.

Initiate Scrollytelling

Beyond a screen reader.
A new way to experience intelligence.

Lexara Air is a next-generation educational platform that transforms complex exam papers into cognitively structured, dyslexia-friendly experiences using Generative AI (Vision + Audio).

The Problem

The Linear Blindspot.

Traditional exam systems assume fast, linear reading — creating insurmountable barriers for dyslexic and cognitively diverse students.

"When the interface itself is a puzzle, the student's true potential remains locked behind a decoding barrier."
The Solution

Cognitive Scaffolding.

Lexara Air restructures information into step-by-step cognitive checkpoints, enabling understanding without cognitive overload.

Restructure

Dynamic Layouts

Simplify

Syntax Leveling

Cognitive Load
+ Accessibility Engine

Dynamically adapts question complexity to match cognitive threshold.

Live Adaptation Active
Exam Item #08

MODERATE SYNTAX DENSITY

Identify the political and cultural capital of the French Republic.

Key terms isolated: Political Capital, French Republic. Synthesis required.

System Architecture.

Modular accessibility engineering for every aspect of the learner experience.

Intelligent PDF Parsing

Vision-based parsing of exam papers. Extracts structured questions from raw documents using proprietary OCR training.

Module LX-1

AI Complexity Analyzer

Detects cognitive load using sentence length, concept density, and vocabulary complexity metrics.

Module LX-2

Dyslexia-Optimized Simplification

Rewrites questions into high-legibility format without answer leakage, using clean vertical layouts.

Module LX-3

Structured Cognitive Steps

Breaks complex problems into discrete logical phases: Identify → Analyze → Solve.

Module LX-4

Lexara Audio Pipeline

AI-powered text-to-speech with controlled pacing and tone variation to support auditory learners.

Module LX-5

Multilingual Engine

Native support for 9+ Indian languages including Hindi, Marathi, Bengali, Tamil, Telugu, and more.

Module LX-6

Dyslexic Typography Engine

OpenDyslexic font toggle with adaptive spacing and line-height controls to eliminate visual stress.

Module LX-7

Cinematic Focus Mode

VisionOS-inspired mode with one-step-at-a-time display and 4-7-8 breathing technique overlays.

Module LX-8

Dyslexic PDF Export

Converts digitized exams into printable, alignment-aligned PDFs with embedded accessibility fonts.

Module LX-9

Seamless Workflow.

From raw document to cognitive empowerment in seconds.

1

Upload Exam

PDF / Image

2

AI Parses

Vision Extraction

3

Select Language

9+ Supported

4

Step Mode

Cognitive Split

5

Audio Guidance

Sync TTS

6

Export PDF

Print Ready

Understanding
over Decoding.

Lexara Air isn't about shortening the path—it's about removing the boulders, so students can walk the path on their own terms.

Making education accessible without lowering standards.

Cognitive Optimization

Designed for how
the brain reads.

Standard Academic Format

Solve the quadratic equation and determine the nature of its roots.

Traditional dense presentation often creates visual crowding, leading to letter swapping and cognitive fatigue for neurodivergent learners.

Accessibility is not about simplifying intelligence.

It's about removing unnecessary cognitive barriers.

Weight Distribution

Heavier bottoms on letters act as a gravity anchor, preventing visual rotation of glyphs like 'p', 'q', 'b', and 'd'.

Vertical Scaffolding

Linear step-by-step breakdown ensures that the eye follows a predictable, non-overwhelming path through complex data.

What I Learned

Designing for Neurodivergence

Deepened my understanding of visual processing in dyslexia. Learned that spacing, line-height, and font-weight are as critical as the actual AI logic in accessibility.

Cognitive Load Optimization

Implemented logic-based complexity detection. This taught me how to scale information density dynamically based on real-time concept analysis.

The Engine Room

Built with a high-performance stack for real-time AI inference and cognitive processing.

Next.js
FastAPI
GPT-4o Vision
PyTorch
gTTS
Tesseract
FAISS
Redis
AWS