To authors considering using Indexia

The ASI evaluated an index taken from the book of an author who publicly acknowledged using Indexia. This method fundamentally undermines the integrity of the review. The ASI was not able to see which parts of the index were Indexia generated and which parts the author edited themselves. The author in question used Indexia to generate a first pass of their index and edited it.

Equating Indexia’s outputs with the edits of an author using our platform is a fundamental methodological flaw.

The ASI, as noted, used an index from an author’s book instead of testing Indexia directly. As such, they were unable to experiment with Indexia’s editing features, which alone would have rectified many of the alleged flaws they document. This oversight is perplexing to us. As early as November of 2025, we invited the AI Committee to test a version of the software free of charge.

Reviewing software without having used the software is an omission, to say the least.

The index that the AI Committee reviewed was among the very first that Indexia ever produced, created on 6 December 2025. In the time since, Indexia developers have shipped 958 updates to the software. These updates were made in response to feedback from the many hundreds of authors we have worked with. Software doesn’t sit still; it improves over time.

The index reviewed by the AI Committee is simply not representative of Indexia’s current capabilities.

This review is a part of a larger body of anti-AI work from the AI Committee. The committee has released a series of white papers on AI and indexing that come to a similarly categorical conclusion: AI is not useful for book indexing. Yet, when we reached out to the AI Committee to request replication data for their results, they refused to share the data. To quote the National Academies of Sciences, Engineering, and Medicine: “complete sharing of data, models, and code when publicly reporting results, are fundamental to research integrity.”

Readers be wary of the reviewer who conceals their methods from scrutiny.

The committee chose to critique an index generated by Indexia and edited by a self-published author. Perhaps this particular index did not live up to the standards of the ASI. However, the index did meet the needs of the author with whom we worked. Is that not sufficient?

This is emblematic of a broader problem. Throughout the review, the AI Committee implies that it is only a professional indexer who can generate an index. “The founders and developers [of Indexia], Ben Vagle and Will Dinneen, are not professional indexers,” the AI Committee writes. This is true. We are academic authors who built Indexia because we understand firsthand how painful and expensive indexing can be.

We do not claim — and never have claimed — that an unedited Indexia index is better than that generated by a true professional indexer. However, hiring a professional indexer is often unaffordable for authors. For example, the chair of the AI Committee’s own indexing rates are “$12–32 per thousand words of text.” Had the author of the reviewed Indexia index hired this professional to index their self-published book, it would have cost them $1,000–$4,000. With Indexia, it cost $99.

Indexia is not intended to replace professional indexers when both the budget and editorial demands of a project call for a professional. However, where the budget doesn’t allow for the creation of a professional index, or where an author, lacking indexing skills, is tasked with creating an index for their own book, Indexia provides a third choice.