Josh SoftCode

Read Better. Understand Faster. Learn Smarter.

Comment

Inspiration

This project was inspired by a common problem in education and daily learning: people often struggle to understand documents, not because they cannot read them, but because the content is too dense or complex. Traditional PDF readers and document viewers only display text without helping the user understand it.

Students especially face this issue when studying handouts or textbooks where concepts are packed into long paragraphs. I wanted to build something that goes beyond reading—something that actually teaches while you read.

What it does

This application is a smart file reader with an integrated AI explanation system.

It:

  • Opens and reads PDF and text documents
  • Automatically detects sections and organizes content
  • Provides a clean, readable interface with adjustable text size and accessibility modes
  • Includes an “Explain Section” AI feature

When a section is explained, the AI:

  1. Extracts key terms
  2. Defines them in very simple language (easy enough for a 10-year-old to understand)
  3. Explains how the ideas connect together
  4. Uses analogies when necessary to improve understanding

How we built it

The system was built in layers:

1. Document processing layer

We built a parser to extract text from files and divide it into logical sections using headings, spacing, and paragraph structure.

2. Reading interface

A clean and accessible UI was created with:

  • Large, readable fonts
  • Text reflow for mobile-style reading
  • High contrast mode
  • Simple navigation between sections

3. AI explanation engine

We designed a structured prompt system that forces the AI to respond in a consistent format:

  • Key terms extraction
  • Simple definitions
  • Concept relationships
  • Analogies

4. Interaction system

Each section includes an “Explain” button that sends the selected text to the AI and displays the structured explanation instantly.

Challenges we ran into

One major challenge was handling unstructured documents. Many PDFs do not have clear formatting, making it difficult to reliably detect sections.

Another challenge was ensuring the AI always explains concepts in a simple and consistent way. Without strict prompting, responses could become too complex or inconsistent.

We also had to carefully balance simplicity in the UI while still keeping the tool powerful enough for advanced users.

Accomplishments that we're proud of

We are proud that we were able to turn a simple file reader into an interactive learning assistant.

Key achievements include:

  • Successfully integrating AI explanations into document reading
  • Creating a structured system for simplifying complex knowledge
  • Designing an accessibility-focused reading experience
  • Making learning more interactive instead of passive reading

Most importantly, we built something that can genuinely help users understand what they read.

What we learned

Through this project, we learned:

  • How important document structure is for understanding content
  • How AI can be used as a teaching tool, not just an answer generator
  • The importance of UX design in accessibility tools
  • How breaking information into steps improves clarity and learning

We also learned how to design prompts that enforce structured, reliable AI output.

What's next for Josh SoftCode

Next, we plan to expand the system into a more powerful learning platform.

Future improvements include:

  • Image and diagram understanding using OCR and vision models
  • Personalized learning modes (beginner, student, revision mode)
  • Offline AI support for accessibility in low-connectivity areas
  • Mobile-first version for wider accessibility

The long-term goal is to evolve this into a full AI learning assistant that makes any document easy to understand for everyone.

Built With

  • chatgpt
  • claud
  • replit
Share this project:

Updates