And the question lands: do I actually need one?

The honest answer isn't a flat yes or no — it depends on your book type, your page count, and who your readers are. But for most nonfiction authors publishing a 100+ page book on Amazon or through any retailer, skipping the index is a quiet mistake that costs you more than you'd expect.

Here's how to decide.

The Types of Nonfiction That Almost Always Need an Index

Not every nonfiction book needs an index, but certain categories effectively require one if you want to be taken seriously by readers and reviewers.

Reference books and guides are the clearest case. If readers will return to your book repeatedly — looking up a specific term, finding a chapter they remember but can't locate, checking a fact — an index is the navigation system that makes the book genuinely useful over time. Without it, readers flip through pages hoping to find something they remember seeing. With it, they go directly to what they need. That difference matters to how readers rate and recommend your book.

Business books, leadership titles, and professional development books are the second clear category. Your reader bought this book because they want to apply the ideas. When they're six months post-reading and trying to locate your three-step framework or the case study about the company they found compelling — they'll use the index. If it doesn't exist, many will simply give up and find another source.

History, biography, and narrative nonfiction sit in the middle. Academic presses typically require indexes for these. Trade publishers usually include them for 200+ page works. As a self-published author, the practical rule is this: if readers will search your book by a person's name, an event, or a specific topic, they expect an index.

When You Can Reasonably Skip It

Not every nonfiction title needs a full back-of-book index. Shorter books — under 100 pages — rarely benefit from one. The effort exceeds the payoff when readers can scan the whole text in under an hour.

Narrative-heavy books where readers engage cover to cover rather than dipping in and out are also candidates for skipping. A memoir told in strict chronological order, a personal essay collection, a short inspirational title — readers follow the through-line rather than searching for specific content.

Ebooks without print editions are a special case. Traditional indexes rely on fixed page numbers from a typeset PDF. Ebook readers can search full text natively, which reduces — but doesn't eliminate — the value of a structured index. If you're publishing print and ebook simultaneously, build the index for print; it adds credibility to both editions.

The Decision Framework: Four Questions

If you're still unsure, run your book through these four questions:

1. Is it 100 or more pages? If yes, you're in index territory. The industry standard — from professional indexers, library catalogers, and major publishers — treats 100 pages as the threshold where a back-of-book index becomes expected.

2. Will readers search it, not just read it? A book someone reads once for inspiration is different from a book someone consults repeatedly for guidance. If your readers are likely to say "I know this book covers X — I just need to find where," an index turns a good book into an indispensable one.

3. Does your book contain proper nouns worth locating? People, organisations, places, frameworks with specific names — these are exactly what indexes surface. If your book references fifteen research studies, five industry leaders, and a dozen named methodologies, an index isn't optional. It's how readers find the thing they half-remember from chapter eight.

4. Would a missing index embarrass you in a review? This is the most direct question. Professional reviewers, librarians, and experienced nonfiction readers notice the absence of an index in books where one is expected. "Well-written but lacks an index" is a real line in real reviews. It's an avoidable criticism.

What Skipping an Index Actually Costs You

The cost isn't just reader experience. It's credibility.

Readers who buy nonfiction with the intention of using it — applying it, referencing it, keeping it on their shelf — use the index to evaluate whether the book is serious. Browsing an index before buying is a legitimate purchase signal. Some readers, particularly those who've been burned by shallow business books, go straight to the index to confirm depth.

There's also the professional review channel. Library acquisitions, academic reading lists, and editorial reviews for trade nonfiction routinely treat an index as a sign of a book produced to professional standards. It signals that the author understood readers would want to find specific content — and made that possible.

Against all of that: professional indexers charge $4–$8 per indexable page, which means a 250-page nonfiction book typically runs $1,000–$2,000. That figure stops a lot of indie authors cold. The timing problem compounds it: indexers require your final, typeset PDF before they can begin, which compresses your production timeline right when it's already tight.

The cost-and-time barrier is real. It's also no longer the only option.

The Bottom Line

If your nonfiction book is 100+ pages, covers a substantive topic that readers will return to, and contains concepts or names worth locating — you need an index. That's most nonfiction books that aren't short, narrative, or purely inspirational reads.

The question isn't whether to have one. It's how to get one efficiently, at a production stage when your timeline and budget are already stretched.

Onomastic was built for exactly this moment in the publishing workflow. Upload your final typeset PDF, and get a CMOS-compliant index — complete entries, subentries, and cross-references — in under two hours. Authors who assumed indexing was out of their budget or timeline find it's neither.

The index your book needs doesn't have to be the thing that delays it.

I did exactly this. And I'm going to walk you through what came back — the actual output, the specific failure modes, and why general-purpose AI isn't the right tool for this job, even in 2026 when it's genuinely excellent at almost everything else.

What I Expected vs. What I Got

The experiment: I pasted three chapters from a 220-page nonfiction business manuscript into ChatGPT-4 and gave it a detailed prompt. "Act as a professional book indexer. Create a back-of-book index for this content. Include main entries, subentries, cross-references, and accurate page numbers. Follow Chicago Manual of Style conventions."

What came back looked credible on first glance. The entries were alphabetized. There were subentries under major concepts. I could even see a few "see also" cross-references. I understand why an author who hadn't compared this against a professional index recently might accept it and move on.

But here's what emerged when I checked it carefully against my actual typeset PDF.

The page numbers were fabricated. Not approximately right — fabricated. ChatGPT had no access to my typeset document. It was estimating page locations based on the sequence of text in the chat window, which has no reliable relationship to where page breaks fall in a formatted InDesign file. When I cross-referenced the output against my final PDF, roughly 60% of the page number entries were wrong — off by anywhere from 2 to 15 pages in either direction.

Large portions of the text went missing. Because of context window limits, ChatGPT wasn't processing my full chapter content — it was working with whatever fit after the system prompt. Concepts introduced mid-chapter or near the end of a section were either absent from the index or represented by a single page reference when they actually appeared on four or five pages throughout the chapter.

The cross-references were inferred, not reasoned. A "see also: leadership, situational" cross-reference looks authoritative on screen. But when I traced it back to the manuscript, the connection was loose at best — the model had associated two terms that appeared near each other, not terms that a skilled indexer reading for meaning would link as genuinely related navigation points.

The subentries were mechanical. A professional indexer groups references to "leadership" under subentries that reflect how readers navigate the concept: leadership, in crisis situations; leadership, building vs. managing teams; leadership, transformational approaches. What I got was a list of page numbers with vague modifiers — the structural appearance of depth without the editorial judgment that makes subentries useful.

The Research Confirms It

This wasn't an isolated experience. A 2025 peer-reviewed study published by Liverpool University Press tested large language models against professionally created book indexes. The results were specific: ChatGPT failed to index 67 out of 319 pages — 21% of the book's total content went completely unrepresented in the generated index. Claude performed better on raw page coverage but both models produced indexes with significant structural deficiencies that would require substantial human correction before they could be used.

The coverage problem isn't a bug that better prompting will fix. It's structural. A general-purpose LLM processing pasted text has no reliable mechanism to know that your chapter on pricing strategy starts on page 134 rather than page 129, because it never encountered the typeset document where that page break actually exists. The pagination is gone the moment you paste text into a chat window.

Professional indexers understand this completely. Their first and non-negotiable requirement is always the final, typeset PDF — not the manuscript, not the Word file, not a page-numbered draft that might shift. They know that even a few formatting changes can alter page breaks across an entire book, which is why they never start until every other production step is locked.

Why the Workaround Doesn't Work

Many authors discover the token limit problem and try to work around it by processing chapters individually: paste Chapter 1, get an index fragment, paste Chapter 2, combine them manually.

This approach compounds the errors rather than solving them.

Index entries that should appear across the full book get siloed by chapter. A recurring framework mentioned in Chapters 2, 5, 7, and 11 shows up as four separate partial entries unless you manually reconcile them — which requires holding the full book's conceptual structure in your head while doing detailed editorial work. Cross-references can only connect concepts within the chunk you're currently processing. Page numbers remain wrong because pasting text into a chat window doesn't restore the layout information that determines where page breaks fall in your formatted file.

The editing time required to fix a chapter-by-chapter ChatGPT index consistently runs 6–10 hours for a 250-page book, according to authors who've documented the process. That is more total time than most DIY indexing approaches, and it produces less reliable output.

What Professional Indexers Actually Charge

Before looking at the alternative, it helps to understand why so many authors turn to AI in the first place.

Professional book indexers charge $2.50–$6.00 per indexable page, according to current industry rate guides. For a 250-page nonfiction manuscript with 210 indexable pages, you're looking at $500–$1,500 before any complexity adjustments. Academic texts, legal references, and technical manuals with dense terminology consistently land at the high end. A serious 300-page reference book can run $1,500–$2,400 for professional indexing.

The timeline compounds the budget concern. Indexers require your final typeset PDF and typically work at 8–10 pages per hour — meaning a 250-page book represents 25–35 hours of professional work. Most quote a two-to-four week turnaround. If your print submission window is ten days away, no budget on earth closes that gap.

These constraints are real and they're not going away. They're exactly why authors are searching for "ChatGPT book index" in the first place.

What a Purpose-Built Tool Does Differently

The key technical difference between a general-purpose AI and a purpose-built indexing tool is where the work starts.

General-purpose AI tools accept text. Purpose-built indexing tools ingest your actual typeset PDF — the identical file your professional indexer would require. They extract text with page-location data from the document structure itself, not from probabilistic inference about where content might fall. That structural difference is what makes accurate page references technically possible in the first place.

From there, these tools take an approach that mirrors the professional workflow rather than replacing it with keyword extraction:

They identify indexable entities in context — people, organizations, concepts, frameworks, and key terms — and present them to you before generating the final index. You can review what the tool found, promote entries that deserve more weight, merge variant phrasings of the same concept, and remove terms that don't need to be indexed. You're exercising editorial judgment over the content, not just accepting AI output.

The tool then handles the structural work: grouping variant phrasings under canonical main entries, generating meaningful subentries (not just page lists), adding cross-references where genuine relationships exist, and producing CMOS-compliant output in two-column format ready for InDesign insertion.

The complete workflow — PDF upload, entity review, index generation, and a thorough human editing pass — runs under two hours for most nonfiction manuscripts. A professional indexer performing the same work clocks 25–35 hours. The output quality is not identical to the best professional indexers working on complex academic texts. But for trade nonfiction, business books, self-help, and most other commercially published manuscripts, it consistently clears the editorial bar that matters: accepted by publishers, useful to readers, professionally formatted.

The Right Tool for the Right Job

The framing that resolves most of the confusion: general-purpose AI and purpose-built indexing tools aren't competing for the same task. They have different designs, different inputs, and different appropriate uses.

ChatGPT and Claude are exceptional for tasks where deep language generation is the primary challenge — drafts, rewrites, research synthesis, promotional copy, outlines, and summaries. They're not designed to process document structure, maintain page-accuracy across 280 continuous pages, or apply the editorial logic that distinguishes a navigable index from a keyword list.

Purpose-built indexing tools are narrow by design. They do one thing: accept your typeset PDF and produce a professionally structured index. That specialization is what makes them technically capable of getting it right.

Here's how to think about which option fits your situation:

Hire a professional indexer if your publisher formally evaluates index quality (academic presses, legal publishers), your contract specifies professional indexer credentials, or you have timeline flexibility and the production budget to accommodate a two-to-four week turnaround. The quality ceiling for complex reference texts is higher than any current tool can reach.

Use a purpose-built AI tool if you're self-published or hybrid-published, your production timeline is tight, your budget needs to stay realistic, or you want to understand what your index will look like before making any financial commitment.

Skip general AI for indexing. Not because these tools are bad — they're extraordinary — but because they're genuinely the wrong tool for this specific task. A hallucinated page number in a back-of-book index isn't a cosmetic issue. It actively misdirects a reader looking for something specific. And fixing six hours of hallucinated output is worse than starting from scratch with a tool designed for the job.

See What Your Index Actually Looks Like

The fastest way to understand the quality difference is to try it with your own manuscript.

Onomastic processes your actual typeset PDF and returns a CMOS-compliant index with accurate page references, hierarchical entries, meaningful subentries, and cross-references built for reader navigation — not keyword extraction dressed up to look like one.

Upload your final PDF. Review the extracted entities. See what the generated index looks like. You'll know within an hour whether the output clears your bar — and you'll have a concrete, real-output basis for comparing it against a professional quote, against your DIY estimate, or against whatever ChatGPT produced when you tried it first.

The experiment I ran with ChatGPT cost several frustrating hours and produced an index I couldn't use. Getting it right took under two hours with a purpose-built tool and passed editorial review without revision requests. That comparison is the one that matters.

This is genuinely good news. Cindex has been the industry standard for professional book indexers for decades. It's robust, precise, and now costs exactly nothing to download.

But "free software" and "the right tool for your project" aren't the same thing. Before you download it and start building your index, there are a few things worth understanding about what Cindex is, who it was designed for, and what it actually requires from you to produce a usable result.

What Cindex Is — and Who It Was Built For

Cindex wasn't designed with authors in mind. It was built for professional book indexers — the trained specialists who make their living creating indexes for publishing houses, university presses, and legal publishers. These are people who have typically spent months learning the craft of indexing, studied standards like the Chicago Manual of Style and NISO guidelines, and built up institutional knowledge about how to group concepts, structure subentries, and create cross-references that actually help readers navigate complex texts.

Cindex gives those professionals a powerful workspace. It handles alphabetization, manages see-also references, formats output to spec, and integrates with typeset PDF workflows. For someone who already knows how to index, it's an excellent tool that's now free. That matters to the professional indexing community enormously.

The question for self-published authors is different: can you use Cindex effectively without that background? The honest answer is: possibly — but probably not in the timeframe or with the ease you're hoping for.

What "Free" Changed — and What It Didn't

The CC0 license means you can download, use, and modify Cindex with no restrictions and no cost. That's a real change. As recently as 2023, Cindex cost several hundred dollars for a license, which put it out of reach for most self-publishing budgets on top of everything else production requires.

What didn't change is the software itself. Cindex is still a professional indexing tool built around a workflow that assumes you know what you're doing. You'll need to:

• Manually read through the typeset PDF and mark every indexable term as you go
• Create and manage your own entry hierarchy — deciding what becomes a main entry versus a subentry
• Build cross-references by hand based on your understanding of the manuscript's conceptual relationships
• Apply CMOS or publisher-specified style rules consistently throughout

The software tracks and formats what you enter. The intellectual work of deciding what to enter, and how to organize it, is entirely yours. That's not a criticism of Cindex — it's the point. The tool is a workspace for a skilled indexer, not an index generator.

The Real Cost: Time and Knowledge

Authors who've documented their experience using Cindex as a DIY solution report one consistent finding: it takes a long time, especially the first time.

A professional indexer working on a 250-page nonfiction book works at roughly 8–10 pages per hour — meaning a full book represents 25–35 hours of focused work. And they come to it with years of training that lets them make fast, correct decisions about entry structure and terminology. A first-time author attempting the same work typically takes significantly longer, because every decision requires deliberation: Is this term worth indexing? Should this go under the main concept or as a subentry? What's the right phrasing for the canonical entry?

The software cost being zero doesn't change the time equation. A 250-page business book indexed in Cindex by a first-time author working carefully is realistically a 40–60 hour project. For most self-published authors, that's a week or more of focused effort during the window right before a print submission deadline — which tends to be the worst possible time to take on a steep learning curve.

There's also the question of output quality. Professional indexes aren't organized keyword lists — they reflect editorial judgment about how readers will actually navigate the text. Developing that judgment takes time and practice. Most first-time authors who use Cindex produce something serviceable, but the gap between a careful first attempt and what an experienced indexer would produce is noticeable to anyone who's seen both.

Cindex vs. AI Indexing Tools in 2026

The relevant comparison for most self-published authors isn't really "Cindex vs. hiring a professional indexer." Professional indexers charge $500–$2,000 for a typical nonfiction book and require a 2–4 week turnaround — a significant budget and timeline commitment many self-published authors can't accommodate. The comparison that actually matters is "Cindex vs. a purpose-built AI indexing tool."

Here's how they differ on the dimensions that matter:

Input requirement: Both require your final typeset PDF — not the manuscript, not the Word file, not a page-numbered draft. The index must be built from the formatted, paginated document where page breaks actually exist. This part is the same.

Who does the intellectual work: In Cindex, you do. You read the text, identify what's indexable, create the structure, and build every entry. In a purpose-built AI tool, the system identifies and surfaces the indexable terms, groups variants under canonical entries, builds an initial hierarchy, and generates subentries and cross-references — then hands you a review layer where you exercise editorial judgment over what it found.

Time investment: Cindex at full professional speed is 25–35 hours for a 250-page book. For a first-timer, realistically double that. AI-assisted workflows consistently run under two hours for the same book, including a thorough human review pass.

Learning curve: Cindex requires you to learn both the software and the craft of indexing. Purpose-built AI tools are designed so you don't need indexing expertise to produce a professional result — the system handles the structural decisions, and you focus on reviewing and refining.

Quality ceiling: Cindex in the hands of a trained professional indexer produces the best possible result. AI tools aren't at the same ceiling for complex academic or legal texts requiring deep specialist knowledge. But for trade nonfiction, business books, self-help, and most commercially published manuscripts, AI-generated indexes consistently clear the professional bar — accepted by publishers, useful to readers, properly formatted.

Total cost: Cindex is now free. AI indexing tools vary in pricing. But the right comparison includes your time. At a conservative $50/hour opportunity cost, 40 hours in Cindex represents $2,000 in time — more than the professional indexer you were trying to avoid.

Who Should Actually Use Cindex

Cindex is genuinely the right choice for specific situations:

Professional indexers who already know the craft and want a free, powerful workspace. The open-source transition is a significant gift to the professional indexing community, and the tool's capabilities are unchanged.

Authors who want to learn indexing as a long-term skill — people planning to publish many books and willing to invest the time to learn the craft properly, treating it as a durable capability rather than a one-time production step.

Publishers handling specialist technical texts — academic manuscripts, legal references, medical texts — where the conceptual complexity genuinely requires deep human engagement with the material that current AI tools don't fully match.

If none of those descriptions fit your situation — if you're a self-published author with one or two books, a timeline measured in weeks, and a need to produce a professional-quality index without becoming an indexing expert — Cindex being free doesn't change your calculus as much as the announcement might suggest.

What Self-Published Authors Actually Need

The news that Cindex is now free is real and worth knowing. It removes a genuine financial barrier for authors willing to invest the time to learn professional indexing software. If that's the path you want to take, opencindex.org has everything you need, and the community maintaining it is active.

But if what you actually need is a professional back-of-book index for a nonfiction manuscript — one that will pass editorial review, help readers navigate your book, and not consume a significant portion of your pre-launch bandwidth — the more relevant question isn't "can I get the software for free?" It's "what's the fastest path to a result I can be confident in?"

Onomastic is a purpose-built AI indexing tool that starts where professional indexers start: your final typeset PDF. Upload your manuscript, review the extracted entities, and produce a CMOS-compliant index with accurate page references, hierarchical entries, meaningful subentries, and cross-references designed for reader navigation — in under two hours.

Cindex being free is good news for the indexing world. It just doesn't change what most self-published authors actually need: a professional result on a realistic timeline, without a 40-hour learning project standing between them and their launch date.

For first-time nonfiction authors, the index is often the most mysterious part of the production process. It's not editing. It's not formatting. It shows up at the back of finished books looking like it generated itself. And once your publisher or production checklist mentions "indexing," the natural first question is: what is actually supposed to go in there?

This is a plain-English answer to that question.

What a Book Index Is (And What It Isn't)

Let's clear up three common misconceptions right away.

A book index is not a table of contents. Your table of contents lists chapters and sections in the order they appear. An index is organised alphabetically and points to specific concepts, names, and terms wherever they appear throughout the book — across chapters, not just within them.

A book index is not a glossary. A glossary defines terms. An index locates them. Your glossary explains what "cognitive load" means; your index tells the reader that "cognitive load" is discussed on pages 14, 47, 89, 92, and 178.

A book index is not a keyword list. This is where a lot of DIY attempts go wrong. A keyword list extracts prominent words. A book index organises ideas in a way that reflects how a reader actually thinks when they're looking for something specific.

What a book index is: a reader navigation system. It's the answer to "I know this book covers X — where exactly is it?" The better the index, the more useful the book becomes as a long-term reference rather than a one-time read.

The Three Building Blocks of Every Index Entry

Professional indexers — and the Chicago Manual of Style, which sets the standard for most trade and academic publishing — recognise three fundamental components of a book index.

Main entries are the primary terms: concepts, proper names, organisations, frameworks, places. Think "leadership," "climate change," "Roosevelt, Theodore," or "agile methodology." These are the headings under which all the relevant page references will cluster.

Subentries are what separate a useful index from a flat one. Rather than listing "leadership, 14, 47, 89, 121, 178" — which tells the reader nothing about which aspect of leadership appears on each page — a well-crafted index breaks that concept into meaningful subdivisions:

leadership — in crisis situations, 47, 89 — building vs. managing teams, 121 — transformational approaches, 178 — common mistakes, 14

Subentries are editorial judgment in action. They require someone — human or purpose-built AI — to actually understand what the text is saying, not just where the word appears.

Cross-references connect related terms that a reader might search under different headings. There are two kinds: See (when a term redirects entirely to another entry — "Roosevelt, Theodore. See presidents, U.S.") and See also (when a term stands on its own but has meaningfully related entries — "leadership. See also management styles").

These three elements — main entry, subentries, cross-references — are what publishers mean when they specify a "professional-quality" index.

What Terms Belong in a Book Index

For first-time authors building or reviewing an index, the question of what to include is often more confusing than the mechanics.

Include:

• Key concepts, ideas, and arguments central to your book's purpose
• Proper names — people, organisations, places, studies, laws
• Recurring frameworks or models you introduce and name
• Terms your target reader would plausibly search for
• Case studies or examples, indexed under both their subject and their topic (a case study on Apple's product strategy should appear under "Apple Inc." and under "product strategy")

Exclude:

• Function words: prepositions, conjunctions, articles ("of," "and," "the")
• Passing mentions — a word that appears once in a transitional sentence doesn't earn an index entry
• Overly generic terms that would generate unhelpfully long page ranges ("book," "author," "idea")
• Chapter titles, which belong in your table of contents, not your index

A useful benchmark: Penn State University Press indexing guidelines suggest that a well-indexed nonfiction book should produce roughly five to six index references per page of text. For a 200-page business book, that's approximately 1,000–1,200 individual references across all entries.

How Genre Changes What You Index

Genre shapes what a useful index looks like — significantly.

Business and self-help books benefit from indexes heavy on frameworks, actionable concepts, and named methodologies. Readers return to these books as references long after the first read. Your index should help them find the specific tool or principle they remember absorbing.

Cookbooks present a layered challenge: you're indexing recipes, ingredients, techniques, and equipment simultaneously. A recipe for Lemon Herb Roast Chicken might legitimately appear under the dish name, under "chicken," under "roasting," and under "lemon" — depending on what a reader might search for.

Memoirs and narrative nonfiction index more lightly than academic works, focusing on proper names, locations, and key events rather than abstract concepts. Many memoir publishers don't require a formal index at all, and a light touch often serves this genre best.

Academic and reference books carry the heaviest indexing expectations. Dense terminology, cross-disciplinary concepts, and reader populations who use these books repeatedly as working references mean indexes that are correspondingly deep and precise. Professional indexers charge $3–$6 per indexable page for standard nonfiction — and up to $8–$12 per page for certified medical or legal indexing, where specialised subject expertise is required.

Print vs. Digital: Does the Format Matter?

For print books, index entries reference specific page numbers — which is precisely why indexing always happens last, after the final typeset PDF is locked and no further layout changes can occur. A single paragraph reflow can shift page breaks across an entire chapter. Build your index from anything other than the final file, and you'll rebuild it later.

For digital-only eBooks, traditional page-referenced indexes rarely translate. Most eReaders use location markers or percentages rather than fixed page numbers. Many ebook publishers skip the back-of-book index entirely and rely on the device's built-in search function.

Hybrid releases — print plus ebook — typically include the full index only in the print edition, with a note in the ebook directing readers to the print version for reference.

The Most Common Mistakes First-Time Authors Make

Indexing too early. The index must be built from your final typeset PDF. If you index a Word document or an earlier layout draft, every subsequent page shift will make those references wrong. Wait until layout is completely locked.

Indexing from the author's vocabulary. Authors naturally notice their own phrasing and terminology. A good index, however, is built for the reader's vocabulary — not the author's. "Brand identity" and "how we want customers to perceive us" may express the same idea in your manuscript. The index entry should use the term readers will actually search for.

Treating it like a word search. A word appearing on a page is not automatically an index-worthy entry. Indexing is editorial work. It requires judgment about what's meaningful, what's searchable, and what will genuinely help a reader navigate the book on a return visit.

If you're a nonfiction author approaching production, the index doesn't have to be the mysterious final step. Understanding what goes into one — main entries, subentries, cross-references, genre-appropriate depth — is the first step to getting it done without the confusion or the sticker shock.

Onomastic was built for exactly this production stage: upload your final typeset PDF, review the extracted entities, and get a CMOS-compliant index with accurate page references, meaningful subentries, and properly reasoned cross-references — in under two hours. See what your index looks like before committing to any approach.

I closed the email. I sat with it for a while.

If you're a nonfiction author trying to figure out what indexing your book will actually cost in 2026 — and whether you have to spend that much — this is the breakdown I needed and couldn't find. No vague ranges, no affiliate disclaimers. Just an honest side-by-side of every realistic option, with real numbers and clear guidance on who each one is right for.

What a Professional Indexer Actually Charges

The short answer: more than most authors expect, with a range wider than most pricing guides admit.

As of 2026, the Society of Indexers (UK) recommends rates of £3.80 per page for non-academic text — roughly $4.80 at current exchange rates. American indexers typically work in the $4–$8 per page range, according to industry pricing guides. Run those numbers on a typical nonfiction book:

That's before complexity adjustments. Academic texts, legal books, and technical manuals with dense terminology typically land at the high end — or above it. A professional index for a serious 300-page reference book can comfortably run $1,500–$2,400.

For context: Reedsy's 2026 publishing cost data puts a full self-publishing production budget at $2,000–$5,000 including editing, cover design, and formatting. Spending $1,000–$2,000 of that on indexing alone — a single component of back matter — is a decision that deserves scrutiny.

There's also the timeline problem. Professional indexers require your final, typeset PDF before they can begin. You can't start the clock until everything else is locked. At an industry-standard pace of 8–10 pages per hour, a 250-page book represents 25–35 hours of work. Most indexers quote two to four weeks minimum, and many have waitlists.

If your print submission window is tight, this timing can be genuinely impossible to accommodate — regardless of what you're willing to pay.

The DIY Option — What It Really Costs You

Two paths exist for DIY indexing: traditional indexing software (Cindex became free in 2023; Macrex and SKY Index remain paid), or Microsoft Word's built-in index tools.

The financial cost is low. Cindex is now a free download. Word is already on your machine. But the real cost is time — and most authors dramatically underestimate it.

Here's what the data shows:

• Professional indexers — people who have done this for years — clear roughly 8–10 pages per hour. A 250-page book takes them 25–35 hours.
• Self-publishing authors report a much wider range. One experienced author documented 5–20 hours on their own projects. Others recommend budgeting three to four full weeks for a first attempt.

The gap reflects the learning curve. Cindex requires training to use effectively; it's built for professional indexers, not for authors who index once every few years. Word's indexing tools are more accessible but more limited — they can build a concordance (a keyword list with page numbers), but they can't generate the hierarchical entries, subentries, cross-references, and double-posting logic that make an index genuinely useful to readers.

The other hidden cost: judgment calls. Every few pages you'll face decisions — does this term get its own entry, or become a subentry under something broader? What do you do with a concept that appears under three different phrasings throughout the manuscript? Do these two tangentially related terms need a see also cross-reference? Professional indexers make these calls intuitively. Authors doing it themselves spend real time on each one.

DIY works well for authors who have indexed before, are indexing a short and conceptually simple book, and have 30+ hours of patience and focus available. For everyone else, the math starts to look less appealing quickly.

Using ChatGPT or Claude: What Actually Happens

The instinct makes sense. AI is embedded in author workflows in 2026, it's fast, and it can process large amounts of text. Why not indexing?

The honest answer is that general-purpose AI tools produce something that looks like an index — and that's the problem.

What they generate is typically keyword extraction with guessed page numbers: a list of terms that appear frequently, with locations assigned based on the model's best approximation of where it encountered them in the text. The structural problems are predictable:

• Hallucinated page numbers: LLMs don't inherently understand pagination. They can guess, but they cannot verify against your actual typeset file.
• Missing cross-references: Linking "revenue recognition" to "GAAP compliance" via a see also requires domain understanding general AI doesn't reliably apply.
• No subentry logic: Grouping twenty page references for "leadership" into meaningful subentries — "leadership, situational," "leadership, transformational," "leadership, in crisis" — requires editorial judgment that keyword extraction doesn't provide.
• Token limits: A full 280-page manuscript is a lot of text. Most general LLMs truncate when processing complete books, which creates gaps and hallucinations in the middle sections.

The editing time required to fix these problems is substantial. Several authors have documented spending 6–10 hours cleaning up a ChatGPT-generated index on a 250-page book. At that point, you've spent more time than many DIY approaches and produced a less reliable result.

This is not a general knock on large language models — they're extraordinary tools for many parts of the writing and publishing process. They're just not built for this specific task.

Purpose-Built AI Indexing Tools: What's Different

This category barely existed three years ago. Today there are several tools built specifically for manuscript indexing that approach the problem differently from general AI: they process the complete uploaded PDF, identify page-accurate locations, recognize indexable entities in context, and produce hierarchical entries that follow CMOS or Chicago Manual conventions.

The practical difference matters. A purpose-built indexing tool:

• Processes your entire typeset PDF without truncation — the same file you'd send to a professional indexer
• Identifies entity page numbers accurately from the document structure, not from probabilistic guessing
• Surfaces people, concepts, organizations, frameworks, and key terms with context around each occurrence
• Groups variants and synonyms (so "machine learning," "ML," and "machine-learning" become a single entry)
• Generates hierarchical entries with subentries and cross-references
• Presents everything for human review before finalizing — so you catch errors before they go to print

The result is a first-pass index that's substantially more accurate than anything a general AI model produces, and it compresses what would be 25–35 professional indexer hours into a workflow that most authors complete in under two hours — including their own review pass.

Tools in this category include Onomastic, IndexStudio, and Indexia. Pricing varies by tool and book length, but typically lands in the $25–$80 per book range — a fraction of professional indexer rates for most manuscripts.

The Full Comparison

Here's every option in the same table:

A few things the table doesn't fully capture:

Price and quality don't scale linearly. A professional indexer charging $1,800 will almost certainly produce a better index than any software on the market today. For academic press publications, legal references, or any book where indexing quality is formally evaluated and part of the author's contractual obligations — that premium is worth paying.

Time is often the binding constraint, not money. If your print submission deadline is ten days away, no amount of money will get a professional indexer to turn around a 250-page manuscript faster than the queue allows. This is where purpose-built AI solves a real problem that money alone cannot.

The "professional quality" bar for most trade books is achievable with AI. For a business book, a self-help title, a parenting guide, a health and wellness book — the practical standard is an index that serves readers well, passes editorial review, meets library and retailer standards, and doesn't embarrass the author. A purpose-built AI tool with a thorough human review pass consistently clears that bar.

Budget reality check: 46% of indie authors earn $100 or less per month from their writing, according to 2026 self-publishing surveys. Spending $1,500 on indexing alone — one line item in a $3,000–$5,000 production budget — is a meaningful financial choice when an alternative exists at $50–$80 and two hours of your time.

Who Should Use Which Option

Here's the clearest way I can frame it:

Hire a professional indexer if you're publishing an academic textbook, legal reference, or technical manual where index quality is formally evaluated — or if your publisher requires professional indexer credentials by contract.

DIY if you've indexed your own books before, your manuscript is under 150 pages, and you have 30+ hours and a high tolerance for tedious precision work.

Skip general AI for indexing specifically. It's the wrong tool for this job, even if it's the right tool for many others in your workflow.

Use a purpose-built AI tool if you need professional-quality output on a real production timeline at a realistic budget — which describes the large majority of self-published nonfiction authors in 2026.

The math has genuinely shifted. A focused 90-minute workflow — upload your PDF, review the extracted entries, make your editorial calls, export — can produce a 300+ entry index that a publisher marks accepted without revision. That's not the exception. That's what purpose-built indexing tools are designed to deliver consistently.

Try Onomastic free and run your manuscript through it. See what it produces. You'll have real, indexed output to compare against every other option — and you'll know within an hour whether the quality clears your bar.

I reached out to three professional indexers. The quotes ranged from \(700 to \)1,200. The timelines ran from two to four weeks. Every one of them needed the final typeset PDF before they could even begin — meaning I couldn't start the clock until everything else was done.

I'm a self-published author on a lean budget. That math wasn't going to work.

What followed was 47 minutes of work, a 14-page index, and a publisher sign-off with no revision requests. Here's exactly what happened — step by step — so you can replicate it.

Why Indexing Is the Production Step Nobody Warns You About

Ask any first-time nonfiction author what they're dreading about production and they'll say editing, formatting, or cover design. Almost none of them will say indexing — until they discover it's mandatory for most nonfiction books sold to libraries, academic institutions, or traditional retail channels.

The economics are sobering. Professional indexers typically charge \(2.50–\)6.00 per indexable page — which, for a 280-page business book, puts you squarely in the \(700–\)1,680 range depending on complexity. And the timeline compounds the pain: at an industry-standard pace of 8–10 pages per hour, a 280-page manuscript represents 28–40 hours of labor. Most trade indexers quote two to three weeks minimum.

For authors publishing with a traditional press, those costs often come out of the author advance. For self-published authors funding production personally, they frequently come as a surprise after everything else has been spent.

I nearly cut the index entirely. I'm glad I didn't.

What a Professional Index Actually Does (and Why It Matters)

Before you assume any tool — AI or otherwise — can shortcut the process, it helps to understand what a good index actually accomplishes. A back-of-book index isn't a glorified keyword list. It's a reader navigation system that:

• Groups related concepts under meaningful headings and subentries

• Captures relationships between ideas that appear in different chapters

• Double-posts key entries so readers find them regardless of how they phrase the search

• Uses cross-references ("see also") to connect related terms

• Follows style conventions — typically CMOS 18th edition — for alphabetization, formatting, and page ranges

This is the detail that matters when evaluating any AI-generated index: not whether it found keywords, but whether it organized them in a way that's genuinely useful to a reader hunting for a specific concept at 11pm the night before a presentation.

My 280-page business book had five interconnected frameworks, four recurring case studies, and a lot of overlapping terminology. It needed proper subentries, not a word dump.

The Workflow: What I Actually Did

I used Onomastic, an AI-powered back-of-book index generator built for authors and publishers. Here's the step-by-step breakdown.

Step 1: Upload the final PDF (2 minutes)

I uploaded the print-ready typeset PDF — 280 pages, standard 6×9 trim. I selected "Business / Self-Help" as the content category so the tool could weight industry-specific terminology appropriately.

One thing worth knowing: upload your final typeset file. If page numbers shift after you generate the index, you'll need to regenerate. I made this mistake with a chapter heading change and spent 20 minutes fixing it. Don't do what I did.

Step 2: Review and curate the extracted entities (8 minutes)

Onomastic surfaces the people, concepts, organizations, frameworks, and key terms it's identified across the manuscript before building the index. You can promote entries (increase their weight), merge duplicate variations, or remove terms you don't want indexed.

I promoted three concepts that were central to my book's core argument — the AI had correctly identified them as frequent, but I wanted to ensure they got full subentry treatment. I also merged two variations of a recurring framework name that appeared under slightly different phrasings in early vs. late chapters.

Step 3: Generate the index (4 minutes)

The full index generated in just over four minutes. Output: 312 main entries with subentries, see also cross-references, and page ranges — delivered in a two-column Word document formatted for InDesign insertion.

Step 4: Human review and editing (33 minutes)

This is where most of my time went, and I'd encourage you not to skip it. I read every single entry. My goals were: (1) verify that subentries were meaningful rather than mechanical, (2) check that cross-references pointed to genuinely related concepts, and (3) add anything the AI had missed.

I added six entries that I knew were important search terms for my target reader. I removed two entries that were too granular. I adjusted phrasing on about a dozen subentries to better reflect how my readers think about these ideas — not just how I wrote them.

Total time: 47 minutes.

The Result: What the Publisher Saw

The final index ran to 14 pages — right in the 4–6% of total page count range that professional indexers target for a business book of this length. It contained:

• 312 main entries spanning concepts, proper names, frameworks, and organizations

• Subentries under every major concept, organized by aspect rather than just listing page numbers

• See also cross-references connecting the book's five core frameworks

• Double-posting of key case studies so readers could find them under both company name and topic

The publisher's editorial contact marked the index accepted without revision. That was the moment I knew the AI output, combined with a focused human review pass, had cleared the professional bar.

What This Means for Your Production Timeline and Budget

To put the numbers in plain terms:

• Professional indexer: \(700–\)1,200, two to three weeks, requires final typeset PDF

• AI indexing workflow: significantly less, 47 minutes, same final PDF requirement

The timeline difference is often what matters most. If your publisher has a 10-day window between final typeset delivery and print submission, a three-week indexing turnaround won't work. I was in exactly that situation.

If your production schedule is tight, your budget is lean, or you're not sure whether the index you'd get from a professional would justify $900 — this is worth trying before you write that check.

Try Onomastic free and generate an index from your manuscript. See what it looks like. The worst outcome is that you learn exactly what your index needs before you hand it to a professional.

That's a mistake. Here's why.

Readers Expect It

Pick up any traditionally published non-fiction book and flip to the back. You'll find an index. Readers of non-fiction — especially academic, professional, and reference works — rely on the index to navigate content efficiently. A missing index signals amateur production quality.

It Boosts Discoverability

A good index doesn't just help readers find what they're looking for — it helps them discover content they didn't know existed. When a reader looks up "market research" and sees a sub-entry for "competitor analysis, 45-48," they've just found bonus value they might have missed.

It Affects Reviews and Sales

Readers notice when an index is missing. Check the one-star reviews on Amazon for any non-fiction book without an index — you'll see complaints. On the flip side, a professional index signals quality and completeness, which leads to better reviews and word-of-mouth recommendations.

It's No Longer Expensive or Painful

Traditionally, professional indexing costs \(500-\)3,000+ and takes weeks. Tools like Onomastic are changing that by using AI to generate professional-quality indexes in minutes, not weeks — at a fraction of the cost.

The Bottom Line

If you're publishing non-fiction without an index, you're leaving money and credibility on the table. Your readers deserve a way to navigate your expertise, and a good index gives them exactly that.

Ready to create your index? Try Onomastic free at https://onomastic.app and see how AI-powered indexing works.

1. Read the entire book. Not skim — read. A good indexer needs to understand the structure, themes, and terminology of your work.

2. Identify indexable concepts. Every subject, name, place, concept, and theme that a reader might look up.

3. Create entries with page references. Each concept gets one or more entries pointing to the relevant pages.

4. Build structure. Subentries for complex topics, cross-references between related concepts, see-also references.

5. Edit and refine. Consolidate similar entries, check consistency, verify page numbers, alphabetize.This process takes a skilled professional 2–4 weeks. It is meticulous, intellectual work.## What Onomastic Does DifferentlyOnomastic uses AI to handle steps 1–4 in minutes. Here is the actual workflow:### Step 1: Upload Your ManuscriptYou upload your book as a PDF, DOCX, or EPUB. These are the formats you already have from your publishing workflow.The system processes your document, identifies the page structure, and prepares the text for analysis.

### Step 2: AI Reads and Analyzes Your BookOur AI model reads your manuscript end to end. It is not keyword matching or simple text extraction — it comprehends the content and identifies:

* Key subjects and concepts — the topics your book is actually about

* People and names — authors, historical figures, case study subjects

* Places and organizations — geographic references, institutions, companies

* Themes and recurring ideas — threads that run through multiple chapters

* Technical terms — domain-specific vocabulary your readers will look upThe AI understands context. It knows that "Apple" in a technology book refers to the company, not the fruit. It recognizes that "the author" in chapter 3 might refer to a different person than "the author" in chapter

7.### Step 3: Draft Index GenerationFrom this analysis, Onomastic generates a structured index following established conventions:

* Chicago Manual of Style (CMOS) formatting — the standard for most book publishing

* ISO 999 compliance — the international standard for index structure

* Proper alphabetization — letter-by-letter or word-by-word, consistently applied* Subentries — multi-level organization for complex topics

* Cross-references — "see" and "see also" entries connecting related concepts

* Page references — accurate pointers tied to your manuscript's paginationThis step takes minutes, not weeks.

### Step 4: You Refine in the EditorThis is the most important step, and it is where you come in.The AI gets you approximately 90% of the way to a finished index. But you know your book better than any algorithm. Our editor lets you:

* Merge similar entries that the AI kept separate

* Split broad entries that should be more specific

* Adjust wording to match how your readers actually think about the topic

* Add cross-references that only a subject expert would know to include

* Remove entries that are technically correct but not useful for your readersMost authors spend 1–2 hours refining. Compare that to 20–40 hours of DIY indexing in Word.

### Step 5: Export and PublishWhen you are satisfied, export your index in your preferred format — ready to drop into your manuscript for publication.And then something important happens.

## What Happens to Your ManuscriptThis is the part most authors care about most, and rightfully so. The moment you export your finished index, Onomastic deletes your manuscript from our servers. Not archived. Not moved to cold storage. Deleted.Here is our privacy commitment in plain terms:

* We never store your manuscript beyond the active indexing session

* We never train our AI models on your content — your words are not used to improve our system

* We never share your manuscript with third parties

* You retain 100% ownership of everything you upload and everything you createWhy? Because your manuscript is your most valuable creative and financial asset. Every other AI tool on the market wants to keep your data. We built Onomastic to forget it.

## Is It Good Enough? The honest answer: it depends on what you compare it to. Compared to a top-tier professional indexer with deep expertise in your subject area? The AI draft will need more refinement. A specialist indexer brings domain knowledge that a general AI model does not have. Compared to no index at all? Dramatically better. A refined AI-generated index adds genuine value to your book.Compared to DIY indexing in Word? More consistent, more comprehensive, and finished in a fraction of the time. Compared to a mid-range freelance indexer? Very competitive — especially after you spend an hour or two refining with your subject expertise. The sweet spot: Onomastic gives you a professional-quality foundation. Your editorial refinement turns it into something that is genuinely yours.

## Who Is This For?

* Self-published authors who have been skipping the index because of cost* Multi-book authors who need efficient indexing at scale

* Authors with tight deadlines who cannot wait 3 weeks for an indexer

* Authors who want control over their index rather than outsourcing it entirely

* Traditional authors whose publishers pass the indexing cost to them## Try It EarlyOnomastic is launching soon. Early waitlist members get founder pricing — the lowest price the tool will ever be offered at.Join the waitlist at onomastic.app Your book. Your index. Your control.