Skip to content

Instantly share code, notes, and snippets.

@zhigang1992
Created January 18, 2026 13:00
Show Gist options
  • Select an option

  • Save zhigang1992/aa25f73b184f6786dd195d9b3399d7f4 to your computer and use it in GitHub Desktop.

Select an option

Save zhigang1992/aa25f73b184f6786dd195d9b3399d7f4 to your computer and use it in GitHub Desktop.
LensED Platform Exploration - AI-powered educational exploration platform

LensED Platform Exploration

URL: https://www.lensed.ai (note: non-www version shows 404) Linear Issue: REI-35


Overview

LensED is an AI-powered educational exploration platform that allows users to explore topics through multiple "lenses" (perspectives). It generates AI-curated content tailored to different educational levels and subject areas.


1. Homepage

Users can explore topics via three input methods:

  • Text input: Type any topic to explore
  • Link import: Paste a URL to analyze
  • Image upload: Drop an image to explore visually

Homepage


2. AI Analysis & Lenses View

When you enter a topic (e.g., "machine learning fundamentals"), the system:

  • Generates an AI image visualization
  • Suggests relevant lenses automatically
  • Provides subtopics to explore
  • Lists "Key Facts" with different perspective tags

Lenses View


3. Frame Settings

Customize the educational context:

  • Level: K-2, 3-5, 6-8, 9-12, Undergrad, Grad/Pro
  • Depth: Controls exploration depth (e.g., "Deep Dive", "Explore")
  • Language: Currently English

Frame Tab

Level Dropdown


4. Lenses System

The core feature - 10 lens categories with 40+ sub-lenses:

Category Description Sub-lenses
Science How things work Physics, Biology, Chemistry, Ecology, Engineering
Arts Creative expression Music, Literature, Cinema, Visual Arts, Performing Arts
History Origins, evolution Ancient, Medieval, Early Modern, Modern, Contemporary
Practice Everyday use Home, Work, Hobby, Survival
Philosophy Meaning, ethics Ethics, Metaphysics, Epistemology, Aesthetics
Design Form, visual choices Visual, Product, UX/Interaction, Craft
Culture People, traditions Tradition, Identity, Global, Subculture
Mathematics Patterns, numbers Geometry, Patterns, Statistics, Algebra, Applied Math
Projects Doable project ideas Film It, Build It, Test It, Design It, Teach It, Code It
Geography Regional perspective US, UK, China, India, Brazil, Other

Lenses Panel


5. Applied Settings

Shows current configuration with "Blend" and "Layer" options for combining lenses.

Applied Tab


6. Focus Mode

The most impressive feature - generates comprehensive educational content:

  • Long-form essay with clickable terms for deeper exploration
  • Connect: Links to experts (Andrew Ng, Fei-Fei Li, Yann LeCun)
  • Watch: YouTube video recommendations (3Blue1Brown, Fireship)
  • Read: Book recommendations with Amazon links
  • Links: Interactive resources (Teachable Machine, Kaggle)
  • Keep exploring: Follow-up questions
  • Chat input: Ask questions about the topic

Focus View


7. Authentication

Simple email/password system for saving and sharing explorations.

Sign In

Sign Up


8. Link Import

Can import external URLs (tested with Wikipedia) and apply lens-based exploration.

Link Import


Technical Observations

  • Stack: React SPA hosted on Vercel
  • AI: Uses edge functions for analysis
  • Image Processing: Has optimization (2224KB → 233KB, 90% reduction)
  • UI: Clean dark theme with auto theme based on time of day

Completeness Assessment

Working Well

  • Core topic exploration flow
  • High-quality AI-generated content
  • Comprehensive lens system
  • Rich Focus mode with curated resources
  • Link import functionality
  • Authentication flow
  • Theme switching

Areas for Development

  • Explorations persistence (requires sign-in)
  • Image upload (not fully tested)
  • Path/history feature
  • Blend vs Layer modes
  • Location-based features
  • Social sharing capabilities

Questions to Clarify

  1. Target audience? (K-12, homeschooling, lifelong learning?)
  2. Teacher/classroom management component planned?
  3. Content moderation approach?
  4. Which AI models are used?
  5. API or integration capability?
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment