Nexory

AI-powered semantic memory engine for screenshots, links, and notes.

0
0
0
JavaScript
public

Nexory

Your memory knows it exists. Nexory helps you find it.

Nexory is an AI-powered content memory engine that helps users retrieve previously saved digital content using natural language and voice search.

People save thousands of screenshots, links, notes, tweets, videos, and random ideas every day — but later struggle to find them again. Nexory solves this by intelligently understanding, organizing, and retrieving saved content without requiring users to manually structure everything themselves.


The Problem

Modern digital life creates information overload.

Users constantly save:

  • screenshots
  • links
  • social media posts
  • notes
  • inspiration
  • ideas
  • references

…but retrieval is broken.

Most people eventually experience:

“I know I saved this somewhere, but I can’t find it anymore.”

Current methods such as:

  • gallery scrolling
  • bookmarks
  • folders
  • WhatsApp self-chat
  • notes apps

do not truly understand the content being stored.


Our Solution

Nexory acts as an intelligent memory layer for digital content.

Users can:

  • import screenshots
  • save links
  • upload notes
  • use share-to-Nexory
  • search naturally using text or voice

Nexory then:

  • extracts meaning from content
  • generates embeddings
  • organizes information intelligently
  • retrieves relevant content semantically

Example:

“Show me that black sneaker I saved last month.”

Instead of manually scrolling through hundreds of screenshots, Nexory intelligently surfaces the correct content instantly.


Core Features

Smart Content Import

  • Screenshot import
  • Manual upload
  • Link saving
  • Notes storage
  • Future browser extension support

AI Understanding

Nexory automatically:

  • extracts text from screenshots
  • understands context
  • creates summaries
  • generates tags and categories
  • builds semantic embeddings

Users can search naturally:

  • “show me the UI design I saved yesterday”
  • “find the photography pricing post”
  • “that tweet about startup funding”

Voice Search + Translation

Users can speak naturally instead of typing.

Voice input is:

  1. converted to text
  2. optionally translated
  3. semantically searched

MVP Goal

The first version of Nexory focuses on delivering one magical experience:

A user uploads or imports screenshots, then searches naturally for something they vaguely remember — and Nexory successfully retrieves it instantly.

Example:

“Show me that black sneaker I saved last month.”

Instead of manually scrolling through hundreds of screenshots, Nexory intelligently surfaces the correct content.


Future Expansion

Future versions of Nexory may include:

  • Browser extension
  • Background screenshot detection
  • Mobile share integration
  • Cross-platform saved-content syncing
  • Context-aware AI memory assistant
  • Smart recommendation memory recall
  • Passive content understanding

Why Nexory Matters

Nexory is not just another storage application.

It is an intelligent retrieval engine for human digital memory.

As digital information consumption grows rapidly, the ability to intelligently retrieve previously seen content becomes more valuable than simply saving it.

Nexory helps users:

  • reduce information overload
  • recover forgotten inspiration
  • organize digital memory effortlessly
  • retrieve meaningful content naturally

Team Nexory

Built as part of:

NSK AI × The Udara Project 2026 — The Build

A remote-first African startup build challenge focused on shipping real products publicly across six weeks.

Team Members

Faithfulness Issijude,
Sara Mekonen &
Bimme Audrey Zun


Repository Status

🚧 Active Development

Current Phase:

  • Week 1 — Problem Validation & System Architecture

Current Focus:

  • User validation interviews
  • MVP definition
  • AI retrieval architecture
  • GitHub setup
  • Technical planning

Architecture Diagram

View Nexory Architecture Diagram


License

MIT License


System Architecture

Nexory follows a retrieval-first AI architecture:

User Content
    ↓
Content Processing
    ↓
OCR / Parsing
    ↓
AI Understanding + Embeddings
    ↓
Vector Storage
    ↓
Semantic Search & Retrieval
    ↓
Relevant Results
v0.3.3[beta]