EverBook Lightkeeper

An AI reading companion that helps readers connect timeless stories with their own lives—and carry one enduring light from every reading experience.

Comment

Inspiration

EverBook is a bilingual digital library created to bring timeless stories back into modern life.

Over the past year, I built EverBook as a non-technical founder working closely with AI across literature, translation, editorial design, product strategy, website creation, international communication, and user testing.

But publishing stories was not enough.

I wanted reading to become personal again—not another summary, not another chatbot, but a quiet encounter between a timeless story and a human life.

That idea became EverBook Lightkeeper.

What it does

EverBook Lightkeeper is an AI reading companion that helps readers connect a timeless story with what they are experiencing in their own lives.

A reader can:

  1. Choose a story.
  2. Share a feeling, question, or personal situation.
  3. Receive a structured reading reflection.
  4. Discover one enduring light from the story.
  5. Take away one small action for real life.
  6. Create a personal “Light Card” that can be saved and revisited.

The experience is designed around four elements:

  • What the story reveals
  • What it may illuminate in the reader’s life
  • One light worth carrying forward
  • One gentle, practical next step

Lightkeeper is not designed to replace reading or human judgment. It is designed to help one story remain alive in one human life.

How we built it

EverBook Lightkeeper was created during OpenAI Build Week as a new interactive experience extending the vision of the existing EverBook project.

GPT-5.6 and Codex were core collaborators throughout the project, supporting product reasoning, interface decisions, implementation, debugging, GitHub integration, Vercel deployment, mobile testing, and documentation.

Codex helped translate the original literary concept into a functional Next.js and TypeScript application, including the story selector, personal reflection workflow, structured result page, downloadable Light Card, responsive interface, and server-side reflection route.

The public demo currently runs in deterministic mock mode so judges can experience the complete flow without providing credentials. The repository also includes an optional server-side OpenAI API path, with all credentials kept out of the browser.

The original EverBook library provides the literary and editorial foundation, while Lightkeeper is the new Build Week application built around that vision.

Challenges

The greatest challenge is not simply generating text. It is creating an experience that is:

  • Thoughtful without pretending to be therapy
  • Personal without becoming intrusive
  • Literary without becoming academic
  • Emotionally meaningful without using empty inspirational language
  • Helpful while still respecting the reader’s own interpretation

Another challenge is building a real product in a very short period of time as a founder without a traditional software development background.

Accomplishments that we're proud of

  • Turning a literary and philosophical idea into a working AI product
  • Designing an experience that connects literature, reflection, and real-life action
  • Building as a solo, non-technical founder with GPT-5.6 and Codex as core creative and development collaborators
  • Extending an existing bilingual digital library with a completely new interactive experience
  • Preserving human editorial judgment throughout the AI workflow

What we learned

AI collaboration is most powerful when the human does not surrender judgment.

Models can help connect knowledge, generate possibilities, and build systems, but the creator must still define the purpose, boundaries, structure, and meaning.

This project also demonstrates that one person can now coordinate work that once required separate teams across product, engineering, editorial, design, translation, and international communication.

What's next for EverBook Lightkeeper

After Build Week, Lightkeeper can expand to include:

  • More EverBook stories
  • Chinese and English experiences
  • Reader-controlled reflection depth
  • Private personal reading archives
  • Shareable Light Cards
  • Reading journeys across multiple stories
  • Carefully designed experiences for schools, libraries, and lifelong readers

The long-term vision is simple:

Help every reader carry one enduring light from every timeless story.

Reading Lights the World.

Built during OpenAI Build Week

EverBook as a literary project existed before Build Week, but EverBook Lightkeeper was created during the submission period as a new companion experience.

The story-selection interface, personal reflection workflow, four-part response structure, downloadable Light Card, optional OpenAI API route, deterministic demo mode, responsive application, GitHub repository, documentation, and Vercel deployment were built during Build Week with GPT-5.6 and Codex.

Built With

Share this project:

Updates