MOBI is dead. Kindle Create can only do so much. Send to Kindle is a reading tool, not a publishing pipeline. We take whatever you already have, in whatever state it is in, and produce the file KDP wants: a validated KPF or EPUB 3 that uploads first time, renders correctly, and does not quietly inflate the delivery fee Amazon subtracts from every sale.
Kindle file conversion is the mechanical work of moving a manuscript out of the format it currently lives in and into a format Amazon's ingestion pipeline accepts and renders predictably. In practice that means a KPF package built in the Kindle toolchain, or a validated EPUB 3 that KDP converts on its own side. Everything else is a detour.
It is a narrower job than formatting, and we would rather say so than sell you something bigger. Conversion asks one question: does this file arrive at Amazon intact, upload without an error banner, and weigh what it ought to weigh. Design decisions, typographic craft and interior styling are a separate conversation.
The reason authors pay for it is that the free routes all leak. Kindle Create refuses anything structurally complicated. Save-as-EPUB carries the mess forward. Third-party converters hand you a file that passes on Amazon and fails everywhere else. Somebody has to check, and that somebody is normally not the software.
MOBI has not been an accepted upload format at KDP since 2021. Files built for it still circulate: on hard drives, in old backlists, inside ARC folders, and in the download libraries of authors who published a decade ago and never looked back. They still open on a Kindle. They will not go up as a new title.
What replaced it is KPF, the Kindle Package Format that Kindle Create and Kindle Previewer produce, alongside plain EPUB 3, which Amazon converts on ingestion. Both work. They behave differently in the previewer, they carry metadata differently, and one of them locks you to Amazon's toolchain. We will tell you which one your book should be.
The native Kindle package. Amazon's own tools build it, Amazon's pipeline reads it, and it gives you the tightest control over how the file lands on device.
The open standard. KDP accepts it, converts it, and you keep a file that also works at Kobo, Apple and every library aggregator you may later want.
Retired for upload, still alive on old hardware. We convert out of it, never into it, unless you specifically need an ARC file for an ancient device.
The conversion is not the upload button. Here is the work that sits between the file you have and the file Amazon accepts.
Before anything is converted we open what you sent and establish what it actually is, which is often not what it is called. A .docx that was pasted out of a PDF behaves nothing like a .docx that was typed. An EPUB exported from InDesign carries different damage from one a converter produced. The conversion route depends entirely on that answer, and getting it wrong costs a rebuild.
We build the package Amazon is currently asking for, not the one it accepted three years ago. Correct mimetype, correct container, correct manifest, correct spine order, and a nav document KDP can read. The upload either goes through without an error banner or we have not finished. There is no partial credit at ingestion.
If your backlist is sitting in MOBI, PRC or AZW3 files with no surviving source, we extract from the file itself. Text, structure, images, notes and metadata are pulled back out, rebuilt as modern markup, and reissued as something you can upload today. Old files are evidence, not obstacles.
Amazon subtracts a delivery charge from every 70 percent royalty sale, calculated on the megabytes your file weighs. A carelessly converted image set can add several megabytes and quietly cost you more each month than the conversion did once. We compress, resize and re-encode until the file is as light as it can be without visible loss, and we tell you the final figure.
Before anything is handed over, the package is put through EPUB 3 conformance checking and opened in Kindle Previewer. Errors are fixed, not documented and passed on. If a converted file cannot pass validation, it is not converted; it is broken, and shipping it just moves the problem to you.
You receive the unpacked markup and stylesheets behind the finished file. That means the next person who touches your book, whether it is us, another supplier, or you at midnight before a promotion, has something to work with. Conversion should not end with you dependent on the converter.
Six conversions, six different starting points. The source file decides the method, and the method decides how long it takes. Open any of them to see what came in and what went out.
There is a correct answer for your book, and it depends on where you intend to sell it. This is the first thing we settle, because it changes the entire build.
The Kindle Package Format is built inside Amazon's own toolchain, which means what you see in Kindle Previewer is very close to what the reader gets. Enhanced typesetting behaves predictably. Chapter styling holds. Upload is direct with nothing left for KDP to reinterpret.
The catch is portability. A KPF is an Amazon artefact. It is not an EPUB, no store outside Amazon will take it, and if you ever go wide you will be starting the file again.
KDP accepts EPUB 3 and converts it on ingestion. The same file, unchanged, is what Kobo, Apple Books, Google Play, Nook and every library aggregator expect. One master, many destinations, and no vendor deciding what your source looks like.
The cost is a small loss of control at the last step, because Amazon runs its own conversion on your file. We build to survive that conversion, and we check the result on device rather than trusting it.
These are the six failures we find in files converted by software with nobody watching. They are boringly consistent.
Different origins, different damage, different method. This is the routing table we work from.
| Element | How we build it | Why it matters |
|---|---|---|
| Word (.docx) | Styles mapped to semantic markup, direct formatting stripped | Word's local overrides do not survive conversion and are the most common cause of ragged output. |
| Text extracted, reading order rebuilt, page furniture removed | A PDF has pages, not structure. Everything a Kindle file needs has to be reconstructed by hand. | |
| InDesign (.idml) | Exported, then rebuilt from the export rather than shipped as-is | The native export carries print-era artefacts and CSS no Kindle device honours. |
| Legacy MOBI / PRC / AZW3 | Unpacked, content recovered, rebuilt as EPUB 3 | Amazon stopped accepting MOBI uploads. The content is fine; the container is obsolete. |
| EPUB 2 | Upgraded to EPUB 3 with an NCX fallback retained | EPUB 2 still ingests, but it forfeits enhanced typesetting and modern navigation. |
| Another retailer's file | Vendor markup and DRM-era wrappers stripped out | Files built for Apple or Kobo carry store-specific tags that Amazon either ignores or trips over. |
| Images | Resized to Kindle's rendered dimensions, recompressed, colour-checked | Every megabyte is charged back to you on every royalty sale for the life of the book. |
| Metadata | Title, contributors, language, identifier and series fields written correctly | Wrong or missing metadata is the quietest reason a file gets rejected or misfiled at KDP. |
Conversion is not finished when the software stops. It is finished when the file has been read on the hardware your buyers own.
Paperwhite and the basic Kindle, because greyscale rendering is where a converted image set shows its weaknesses first. Anything that relied on colour to be legible, such as a chart with a red line and a green line, arrives here as two identical grey lines. We find that before your reader does.
A different rendering engine entirely. Files that converted acceptably for e-ink can show compression artefacts, banding in gradients or oversaturated photographs once they are backlit. Both surfaces get checked, because neither one predicts the other.
The screen most Kindle reading actually happens on now, and the one where a converted file with hard-coded pixel widths collapses first. Narrow screens punish anything the converter left hard-coded.
Previewer is where Amazon's own conversion runs, so it tells us what KDP will do to an EPUB on ingestion. Useful and not sufficient. We treat it as the first check, never the last one.
We put the finished file through the actual upload path in a draft, watch for the error banner, and read the conversion report Amazon returns. Finding out at that moment is free. Finding out on launch day is not.
Someone opens the source and the converted file side by side and reads. Chapter counts, image counts, note counts and the last line of the book. Software does not notice when chapter eleven has vanished; a person does.
Named files, in a folder, with a note telling you which one to upload. Nothing mysterious.
The file you drag into the upload box. Built in the Kindle toolchain, carrying your metadata and your cover, and already previewed on our side. If you chose the KPF route, this is the only file you need to publish, and it is the last one anybody has to interpret.
Your portable master. It passes validation, it ingests at KDP, and it is the file every other retailer and library platform will ask for the day you decide Amazon is not the only shop in the world. Keep it somewhere you will find it in five years.
Optional and increasingly rare. Some review teams and older devices still want one, and old habits persist inside ARC groups. We will build it, and we will also tell you honestly if you do not need it.
A short document showing the final file weight, what it was before we started, and what Amazon will therefore deduct per sale. Most authors have never seen this number. It is a real cost, and it belongs on paper.
The XHTML, the CSS and the image assets, unzipped and readable. This is the difference between owning your book and renting it from whoever built the file.
Which file goes where, which KDP fields to fill, what to expect in the previewer, and the two settings authors routinely get wrong at the last screen. Written in plain English, not in support-article language.
Five stages. The first one occasionally ends with us telling you not to buy anything.
We open what you have, identify what it really is, and tell you which conversion route it needs and why. If a free tool would genuinely do the job, we will say that.
The content comes out of its old container. Encoding damage, broken links, phantom breaks and vendor markup are cleared before anything is rebuilt.
Structure, navigation, metadata and images are assembled into a KPF or EPUB 3 that matches Amazon's current ingestion requirements.
File size is driven down, validation is run, and the file is opened on hardware and in Previewer. A failure here goes back to stage three.
Files, source, upload note and the delivery-fee figure. Then two revision rounds, with no argument about whose fault anything was.
Conversions are quoted on length, since reading, checking and rebuilding all scale with how much book there is. Your source format and image volume adjust the figure, and the delivery window moves with your selections.
Required steps are marked with *. Optional extras can be removed from the summary with the cross icon.
This decides how the file has to be rebuilt, and therefore what it costs.
Reading, checking and rebuilding all scale with length, so length is the measure.
Where the book is coming from. This is the single biggest factor in how hard the conversion is.
Tables, notes and verse all survive conversion badly. Tell us now, not on delivery day.
Image volume drives the rebuild time and the per-megabyte charge Amazon takes off each sale.
The base turnaround is derived from your selections and updates on screen as you go.
Two rounds are included. Add more if an editor or co-author is reviewing alongside you.
Editorial work, proofreading, cover artwork, ISBN purchase, advertising spend and any subscription tools sit outside this service. We convert the file you send. We do not rewrite the book inside it.
The software is not lying to you. It is simply not looking, and neither is anybody else until a reader does.
| What you are comparing | Free Tool / Auto-Convert | Book Formatting Experts |
|---|---|---|
| Legacy MOBI backlist | ✕ Refused at upload, because Amazon no longer accepts the format at all. | ✓ Unpacked, recovered and reissued as an EPUB 3 you can publish today. |
| File weight | ✕ Whatever falls out of the converter, charged back to you on every sale forever. | ✓ Driven down deliberately, with the per-sale delivery cost written on paper before you publish. |
| Reading order | ✕ Sorted by filename, which is not the order you wrote the book in. | ✓ Spine order verified chapter by chapter against the original manuscript. |
| Character encoding | ✕ Accents and quotation marks silently replaced with boxes somewhere past chapter nine. | ✓ Swept across the whole file, repaired, and read back to confirm. |
| Kindle Create limits | ✕ Silently refuses complex structure, then produces something simpler than your book. | ✓ Structure rebuilt to survive, rather than trimmed until the tool stops complaining. |
| If the upload is rejected | ✕ You are alone with an error code and a support forum. | ✓ It is our problem. We fix it, and it does not consume a revision round. |
The source file changes the job far more than the genre does, but some kinds of book bring their own specific trouble.
Published years ago, converted by somebody who has since disappeared, and now living only as a MOBI on a hard drive. We open the compiled file, pull the text back out, and rebuild it into something you can upload today. The book survives; only the container was obsolete.
Three hundred footnotes, a bibliography, cross-references and tables. This is where converters do their worst work, because every one of those elements is a link that has to survive a container change. We rebuild them rather than hoping they carry across.
Sixty photographs at full camera resolution, converted straight into the file, and a delivery fee now eating a visible slice of every sale. The pictures stay. The megabytes do not, and the difference is money each month rather than a one-off.
Twelve titles, four different suppliers, three file formats and no consistency at all. Converted together, weighed together, and reissued so that book nine no longer looks like a different publisher made it.
Files built for Apple, Kobo or Nook carry store-specific markup that Amazon either ignores or stumbles over. The wrapper comes off, the content is rebuilt for KDP, and your original file stays intact for the store you are leaving.
Line breaks are the poem. A converter treats them as accidents of layout and quietly turns your stanzas into prose. Verse has to be rebuilt with the breaks protected, or the book has not survived conversion in any meaningful sense.
Any competent developer can run a conversion. The software is free, it is fast, and it produces a file within a minute. What it will not do is open the result and read it, and that is the entire reason authors end up paying twice.
So the value here is not the conversion. It is the person who notices that chapter eleven is missing, that the ellipses have turned into boxes, that the file weighs eleven megabytes when it should weigh three, and that the spine puts the acknowledgements before the prologue.
Chapter count, image count, note count and word count, reconciled against the source. A conversion that drops content silently is worse than one that fails loudly.
No converter watermarks, no stray vendor metadata, no duplicate cover, no empty content files padding out the spine.
File weight is a recurring charge, not a one-off. It is checked, driven down, and reported to you in dollars per sale.
This is the actual list. A file does not ship until every line on it is ticked.
Six situations that bring authors to this page. If one of them is yours, the diagnosis is free.
You published before 2021, the files still work on your own Kindle, and you have just discovered that KDP will not take them for a new edition or a relaunch. The content is recoverable. It is sitting inside those files, and pulling it back out is a solved problem rather than a rewrite.
It will not take your structure, or it flattens something you needed, or it simply stops. Kindle Create is a decent tool with a hard ceiling, and complex nonfiction, academic work and anything with real hierarchy tends to hit that ceiling early. Beyond it, the file has to be built rather than generated.
An exclusive deal ended, or you are consolidating, and the file you have was built for someone else's store. It carries their markup and their assumptions. Amazon will either ignore that or trip over it, and neither outcome is one you want to discover at launch.
You looked at a KDP report and noticed the deduction. On an image-heavy book at the 70 percent rate, that deduction can run to a meaningful fraction of the sale, on every single copy, for as long as the book exists. It is usually fixable in an afternoon.
The freelancer is gone, the agency folded, and you have a finished eBook and no source. We can work from the compiled file. You should still take the source this time, from whoever builds it.
You have written the book, you do not want a design conversation, and you would like somebody competent to hand you the thing Amazon accepts. That is a legitimate purchase, and it is exactly what this page sells.
Three commitments in writing, and one thing we will not pretend about.
If KDP rejects the file we deliver, that is our problem to solve, not yours to research. We will not send you to a help article, we will not blame your source file, and the fix does not come out of your revision rounds. A file that Amazon will not accept is not a delivery.
Chapters, images, notes and words are counted before and after. If something genuinely cannot survive the conversion, and occasionally something cannot, you will be told before delivery and offered the alternatives. You will never find out from a reader.
If your file needs a free tool and ten minutes rather than a conversion service, we will say so and lose the sale. That answer has cost us money and earned us most of our repeat clients, which is a trade we are content with.
Three principles that settle every judgement call, and one always comes up.
The converter has no opinion about whether it succeeded. It reports success as long as it did not crash. Every file that comes out of a machine is treated as a suspect until a person has read it against the original.
A print file is heavy once. A Kindle file is heavy on every sale you will ever make. Megabytes are not a technical detail here, they are a line item, and we treat them like one.
Even when a client only wants Amazon, we build the open standard underneath. Exclusivity ends. Stores change. Nobody has ever regretted owning a validated EPUB 3, and plenty of people have regretted owning only a KPF.
Three things worth keeping, even if the rest of this page leaves you.
What the file is called and what the file is are frequently different things. Every wasted conversion begins with somebody skipping this step.
Translation carries the damage forward in a new wrapper. Rebuilding removes it. The second one takes longer and it is the only one that works.
A file that validates can still be wrong. Check what it costs per sale, then open it and read the book on a device that a customer actually owns.
Rated 4.9 out of 5 across 158 reviewed projects.
I had already paid someone else to format this book once. The chapter headings shifted every time a reader changed the font size and I only found out from a review. This team rebuilt the interior from the structure up and sent me photographs from four real devices before I uploaded anything. First launch day I have not spent refreshing the reviews in a panic.
What sold me was the audit. They came back and told me two chapters had been pasted in from another document and were carrying invisible styles, and that fixing that first would save me money later. Nobody had mentioned it in three years of self-publishing. The proof needed no corrections at all.
My book has footnotes, tables and two appendices, and every quote I got either ignored that or doubled the price because of it. Here the pricing was laid out step by step and I could see what each element cost before I committed. It arrived early and the notes link both ways, which nobody else had even offered.
The proof came back from the printer with the margins exactly where they should be, the running heads correct, and not one chapter opening on a left-hand page. This is my fourth book and the first time I have not had to pay for a second proof copy.
English is my second language and I was nervous about the back and forth. Everything was explained plainly, and when I asked the same question twice they answered as if it were the first time. The revisions were turned round inside two days and I was never made to feel like a nuisance.
I write cookbooks, so the interior is the product. Ingredients have to sit beside the method or the recipe stops working, and the last person I hired reflowed the whole thing into nonsense. These people understood the problem before I had finished describing it and the finished layout is genuinely beautiful.
Straightforward from beginning to end. A fixed price, a delivery window they actually hit, and files that went up to KDP and IngramSpark without a single rejection. I have since sent them two more titles and recommended them to my writing group without hesitation.
My memoir has photographs and I had resigned myself to them looking muddy. They optimised every image, told me honestly which two were too low-resolution to save, and suggested crops instead of shrugging. That candour is worth more to me than the discount I was chasing elsewhere.
I sent them a 480-page reference book with a nightmare of cross-references and a table on nearly every spread. They quoted it accurately, they did not come back later asking for more money, and the index still works. Any author will tell you what that is worth.
The questions authors actually ask us, answered without sales language.
Conversion is the mechanical job of getting your book out of the format it is in and into the format Amazon accepts, which today means a KPF package or a validated EPUB 3. Formatting is the design and typographic work that decides how the interior looks and behaves. They overlap, and they are not the same purchase. If your book is already well built and you simply need it in the right container for KDP, conversion is what you want, and you should not pay for anything larger.
No. Amazon stopped accepting MOBI as an upload format in 2021. The files still open on Kindle hardware and they still sit in author libraries and ARC folders, which is why the confusion persists, but KDP will not take one for a new title or a new edition. If your backlist is in MOBI, the content is entirely recoverable. We unpack the compiled file, pull out the text, structure, images and metadata, and rebuild it as an EPUB 3 you can publish.
KPF stands for Kindle Package Format. It is the file that Kindle Create and Kindle Previewer produce, and it is what Amazon's ingestion pipeline reads most directly. Its advantage is control: what you see in Previewer is very close to what lands on the device, and KDP has nothing left to reinterpret. Its disadvantage is that it is an Amazon artefact. No other retailer will take it, so if you might ever go wide, you want an EPUB 3 as well.
If Amazon is your only channel and you want the tightest control over rendering, KPF. If there is any chance you will sell at Kobo, Apple, Google Play or through a library aggregator, EPUB 3, because that is the same file all of them want and KDP converts it happily on its side. We build the EPUB either way, because the cost of having it and not needing it is nothing, and the cost of needing it and not having it is a second project.
Often you can, and we will tell you so if that is the honest answer. Kindle Create is a competent tool with a firm ceiling. It handles straightforward prose well. It struggles with deep heading hierarchies, complex tables, extensive footnotes, verse, multi-level lists and anything where the structure of the book is doing real work. When it hits that ceiling it does not warn you loudly, it simply produces something simpler than the book you wrote.
It does not, and this catches people out. Send to Kindle is a delivery tool for getting a document onto your own device, or onto a reviewer's device, so it can be read. It performs a rough conversion on the fly and the result is nobody's idea of a finished product. It creates no listing, no royalty, no store presence and no control over what the file looks like. It is useful for proofing on hardware and for sending ARCs. It is not a route to market.
On the 70 percent royalty option, Amazon deducts a per-megabyte delivery charge from every sale, calculated on the size of your finished file. For a plain novel of two or three megabytes, this is small change. For an illustrated book, a photo-heavy memoir or a cookbook converted carelessly, it can run into a real fraction of every single sale, forever, on every copy. We record the weight of your file before and after, and tell you what the difference is worth per sale.
Yes, and there are two numbers worth knowing. The upload ceiling for a Kindle eBook is 650MB, which almost nobody reaches. The number that matters far more is 50MB, which is the point beyond which Amazon's own tooling and several of its delivery paths start behaving badly, and beyond which the delivery-fee arithmetic becomes genuinely painful. In practice we aim to land most illustrated books far below that and most prose books under five megabytes.
Yes, and it is the hardest conversion there is, which is why it costs more. A PDF has no structure. It has pages, positioned text fragments and a visual arrangement that a human reads as a book. Everything a Kindle file needs, meaning reading order, paragraphs, headings, and notes attached to the right sentences, has to be reconstructed. Anyone quoting you the same price for a PDF as for a Word file has not opened the PDF.
It helps, and it is not the shortcut people expect. The InDesign export produces an EPUB packed with print-era assumptions, absolute positioning and CSS that no Kindle device honours. We would still rather have the IDML than the PDF, because the text and the styles are intact and the structure is recoverable, but the export is a starting point rather than a deliverable.
Not one. Your text is your text. We repair encoding damage, which means turning a mangled character back into the accent or quotation mark it was supposed to be, and we will point out obvious typos we trip over, but we do not correct, rewrite or improve your prose. If we think a manuscript needs an editor rather than a converter, we will say so before you pay us anything.
Yes, and backlist migration is most of what pays our rent. Backlists usually arrive as an archaeological dig: three formats, four suppliers, no source files and a decade of drift in the styling. We convert them as a set rather than as unrelated jobs, so the reissued books share a consistent interior and consistent metadata. Volume pricing applies from the second title onwards.
It can, once the store-specific parts come out. Files prepared for Apple, Kobo or Nook carry vendor markup, display options and occasionally proprietary extensions that Amazon either ignores or handles unpredictably. We strip the wrapper, rebuild the package for KDP ingestion, and leave your original file untouched so that the store you are leaving continues to work while the transition happens.
Each one is resized to the dimensions a Kindle actually renders, recompressed to the smallest weight that shows no visible loss, and then checked twice: once in colour on a backlit screen, and once in greyscale on e-ink. Charts that rely on colour to be readable get flagged, because on a Paperwhite a red line and a green line are the same line. The result is a lighter file, which is a cheaper file.
Yes, and we would rather you insisted on it everywhere. You get the unpacked markup, the stylesheets and the image assets. With those, any competent formatter on earth can pick your book up and carry on. Without them, you are tied to whoever built the file, and that is a bad position to be in the week before a launch.
The broader eBook conversion service covers every destination: Apple, Kobo, Google Play, libraries, the lot. This page is about Amazon specifically, which has its own ingestion pipeline, its own retired formats, its own file weight economics and its own toolchain. If Amazon is where the difficulty is, start here. If your problem is a whole catalogue going everywhere, the broader service is the better fit.
Yes. Send us the file and the exact error message Amazon returned, because that message narrows the diagnosis considerably. Most rejections come down to a small number of causes: an unsupported format, a broken package structure, a cover that fails the specification, or metadata that conflicts with an existing listing. We tell you what it is before we quote.
Yes, when it is built properly. A modern EPUB 3 with an NCX fallback in place renders on old hardware as well as new, and Amazon's own conversion handles the backward compatibility on ingestion. What breaks on older devices is usually not the format but the assumptions: fixed pixel widths, fonts that were never embedded, and images sized for a screen that did not exist when the reader bought their Kindle.
The calculator gives you a live window as you build the quote, and it is calculated rather than optimistic. A short prose title converted from Word can move quickly. A PDF rebuild of a hundred-thousand-word academic book cannot, and anybody who promises otherwise is planning to skip the checking. The faster tiers are costed against your own calculated window, and the figure is on screen before you commit to anything.
The file, exactly as it is, including the broken one. Do not tidy it first, do not re-export it, and do not run it through another converter to see if that helps, because that usually adds a layer of damage on top of the one we were going to fix. Send the worst version you own and tell us what happened to it. The diagnosis costs nothing.
Put the file through the calculator and the price and the turnaround are both fixed before you spend anything. If what you have is a mess, that is not a reason to hesitate. It is the reason this service exists.