About

ClearShelf is a frontend-only, dependency-free, AI-built personal editorial project: a lightweight book site built around simple shelves, visible judgment, and visible incentives.

It is not a complete catalog of every book. I decide which books, themes, and shelf ideas ultimately end up on the site, and some shelves begin with a few books I want to test, compare, or build outward from.

That human guidance is part of the point. ClearShelf does not pretend selection is neutral. The aim is to make editorial judgment legible instead of hiding it behind vague authority or automated mystique.

The site is frontend only. It is built as static HTML, CSS, and JavaScript, generated at build time and deployed without a custom backend, database, user accounts, analytics package, or app framework.

The codebase intentionally avoids external packages for the site itself. The public pages are generated from local source files and plain JavaScript rather than a framework stack.

How Books Get Here

ClearShelf should not be read as a claim that I have personally read every book on the site cover to cover. I have not, and part of the reason for using an AI Librarian at all is that no human can realistically read everything worth considering. The point is to widen the field of evaluation without pretending that one person has read everything.

Some books are added because I know them firsthand, but others are added because they fit a shelf I want to build, because they are important comparison points, or because they seem worth evaluating inside the ClearShelf framework. That means the site mixes firsthand reading judgment with structured evaluation, research, metadata review, and rubric-based comparison. A book's presence on ClearShelf means it entered the editorial process, not automatically that it came from my own complete reading history.

But the catalog and shelves are not built only by hand. The Librarian also plays a curation role by helping broaden candidate lists, round out shelves, and surface books that fit the theme and rubric once the process is underway.

That means selection is partly human-guided and partly Librarian-guided. Once books are under consideration, the published ClearShelf rubric is used to score fit and order shelves, rather than having each title hand-tuned ad hoc.

The site makes no claim to universal objectivity, but it does aim for visible standards and consistent treatment once a book enters the process.

The long-term goal is fully autonomous curation and evaluation within the ClearShelf framework. ClearShelf is not fully there yet, and especially early on I still make some decisions about what gets added and how shelves get started.

How The Site Is Made

ClearShelf is human guided but built almost entirely with AI assistance. ChatGPT is used in the writing, coding, structuring, and editorial workflow that produces the site and its content.

There is not a live on-page chatbot or hidden model making runtime decisions while you browse. The AI is used in the creation process, and the final published pages are static.

The scoring rubric is public and can be read and audited at The Librarian AI. In that sense, part of the project is deliberately "proof of prompt": the methodology is visible instead of hidden, even when the upstream catalog and shelf ideas are occasionally personally curated.

When ClearShelf uses generated social post images, those images may be created with Grok image tools and then reviewed before publishing. That image workflow is separate from the website itself and is not required to browse the site.

Privacy And Incentives

As an Amazon Associate I earn from qualifying purchases.

ClearShelf does not set on-site cookies or run tracking analytics. If you click an affiliate link and leave for Amazon, Amazon may use its own cookies and tracking under its own policies.

ClearShelf is also a small family project that I am building with my children. In some cases books may be added to the site where we have a strong interest in them because they were written, edited, published, or otherwise connected to a family member or friend. In those cases the book will be marked as having an additional connection to ClearShelf creators.

The goal is to keep the incentives visible: no hidden ranking marketplace, no account system, no engagement trap, and no mystery about how the pages are put together.