Fullstack GraphQL: an open source book built from markdown chapters

The repository is a manuscript, not a GraphQL stack

Fullstack GraphQL is described in the repository as an open source book about full stack GraphQL development with React and Node, and the short description in the manifest repeats the same scope. That distinction matters before anything else, because a repository named after a stack usually contains that stack. This one does not. The top level entries are the things a book needs: a manuscript folder, a book.toml, a styles folder, a LICENSE, a package.json and GitHub workflow configuration. There is no src directory, no server entry point and no test suite.

So there is nothing here to npm install and start. What you get instead is reading material, hosted for free at graphqlcollege.github.io/fullstack-graphql, and a writing pipeline that lets anyone contribute to it. The primary language recorded for the repository is CSS, which fits a site built around a custom theme rather than an application.

That makes the project a good fit for a developer who wants the shape of a GraphQL application explained in sequence, and a poor fit for a team looking for a server implementation to put behind a production endpoint.

Book.txt and manuscript/SUMMARY.md decide what a reader sees

The build structure is unusually explicit for a book. Each chapter is written in a markdown file inside the manuscript folder, and every image the book uses lives in manuscript/images. The order of the chapters is not inferred from filenames or directory listing. It is determined by the Book.txt file, and manuscript/SUMMARY.md defines the table of contents for the site.

A third file, Sample.txt, defines the content of the book's sample chapter, which the site exposes as a preview of the writing style. The naming suggests a reusable template for the project layout, since the book can be forked and re-skinned without rewriting the pipeline.

The practical upshot for a reader is that the site you visit is a rendering of a git repository, and for a contributor it is that chapter ordering and the table of contents are two separate places to touch. A pull request that adds a chapter markdown file without updating the table of contents produces a chapter the site may not list.

mdBook turns every push to master into a new online book

The website is built with mdBook, the Rust project's book generator, and every push to the master branch deploys the result to GitHub Pages. There is no release process involved, no version tag and no build artifact you download. The book is a branch that keeps moving.

The project also documents a local preview: the command `mdbook serve --open`, run from the repository root, starts the mdBook development server and opens it in your browser, so you can read the book, check how a change to a chapter renders, and confirm that images in manuscript/images resolve before you open a pull request. The book.toml file holds the site configuration, and the styles folder holds the custom CSS theme.

For a reader this is a pleasant arrangement, because the URL never goes stale and there is no version pinning to think about. For someone who wants to quote a specific passage years from now, the absence of tagged versions is a genuine drawback, since the text at a given URL can change without notice.

The npm manifest is a label, not an installable product

The presence of a package.json in a book repository usually signals a JavaScript toolchain. Here the file carries metadata and nothing else. It has no dependencies, no scripts, no main entry and no bin, so npm install produces no runtime, and npm publish of this file would give you an empty package.

{
  "name": "fullstack-graphql",
  "description": "Book about fullstack GraphQL development with React and Node",
  "homepage": "https://graphqlcollege.github.io/fullstack-graphql/",
  "author": {
    "email": "[email protected]",
    "name": "Julian Mayorga"
  },
  "license": "CC-BY-NC-SA-4.0"
}

Two things are worth noticing. The license field names the Creative Commons identifier as a machine readable value, which tells you the intent of the project before you open the LICENSE file. And the author field identifies a single person, Julian Mayorga, which is consistent with a personal project rather than a team maintained course.

The manifest is not a build configuration. If you want to run the book yourself rather than read the hosted one, you need mdBook installed locally, and that is the only real dependency in the workflow.

New chapters need an issue before a pull request

The contribution rules are unusually firm about chapter proposals. The project asks you to file an issue to suggest new chapters, and it asks you not to open a pull request with a new chapter before that discussion, because the chapter may not be accepted. The wording is friendly but the requirement is real: a chapter pull request opened first is not the expected path.

Code fixes take a different path and need no such approval. You open a pull request that explains the fix, and typos are handled through GitHub issues rather than pull requests at all. Questions about the book, or about GraphQL in general, are also routed to issues.

This workflow suits a project whose scope is set by one author. It also tells you something about what to expect from the book: new topics are added deliberately and slowly, not in response to demand, so the table of contents will not track every GraphQL feature that appears in the ecosystem.

The non commercial clause is the real licence constraint

The book is licensed under Creative Commons Attribution Non Commercial Share Alike 4.0 International, and the README displays the CC, Attribution, NC and SA badges rather than an MIT or Apache header. The NC part is the one that decides who can use it. Non commercial use covers personal study, internal learning and nonprofit work. It does not cover shipping the text inside a product you sell, or republishing it as part of a paid course.

The ShareAlike part applies when you adapt and redistribute. A modified version has to carry the same licence, so a fork stays non commercial. Attribution is required either way, and the standard form is to credit Julian Mayorga and link to the original.

For an individual developer this is close to a non issue. For a company producing training material, a documentation site, or a course that mentions the book, the licence has to be read before you copy anything. The repository is described as Other by GitHub because a Creative Commons licence does not fit the categories the platform recognises, which is a sign to open the LICENSE file rather than assume a permissive default.

One branch carries the whole book, with no releases to track

Maintenance information for this repository is thin, and it is worth being precise about what exists. The repository is not archived, and the last push was on 2026-10-07. There are no GitHub releases, so there is no version history to pin, no changelog to read and no upgrade path for a reader, because there is nothing to upgrade.

For contributors the practical implications are different. Every change lands on master and every push deploys the site, so a broken markdown file or a bad image path is visible on the public site shortly after merge. The GitHub workflows folder suggests that publishing is automated rather than manual, though the workflow files themselves are not part of what the documentation describes.

The absence of releases also means the repository cannot tell you how far the text has moved from any earlier version you read. If a course syllabus points at a specific chapter, check the chapter text on the site rather than trusting that the passage is still there in the same form.

Editorial conclusion

Adopt Fullstack GraphQL if you want a free, current, read online reference for pairing a GraphQL API with a React front end, and you are learning rather than shipping a product from it. Do not adopt it if you need a runnable GraphQL server, a library to depend on, or content you can use inside a paid course or product, because the CC BY-NC-SA 4.0 licence blocks commercial reuse. Before you start, check the table of contents in manuscript/SUMMARY.md against the topics you need, since chapter order is fixed by Book.txt and the repository publishes no releases to track.

Frequently asked questions

Official sources

1. GraphQLCollege/fullstack-graphql on GitHub
2. Issues
3. Project website
4. README