EPUB to PDF Converter - Ebook to PDF Online

Convert EPUB to PDF online, along with AZW3, MOBI, FB2, CBZ, LIT and PDB files. Upload one ebook or a whole folder of them, choose a page size, and get typeset PDFs back. Powered by calibre, with no install, no account on some ebook forum and no desktop software.

An EPUB is reflowable: the text has no pages until a reader gives it a screen. A PDF is fixed: every line break is decided once, at conversion time, and never moves again. So this is not a lossless transform, it is a typesetting job, and the settings below are the ones that decide how the result reads. You need a PDF when something downstream demands one - a printer, a university submission portal, an e-reader that refuses EPUB, a colleague with Acrobat and nothing else.

• 📄 A real typesetting pass, not a screenshot - selectable text, working links, a PDF outline and an embedded cover
• 📐 Nine page sizes - A4 and US Letter for printing, A5 and A6 for something that reads like a paperback rather than an office memo
• 🔠 Font, size and margins are yours - serif or sans body text, a base font size in points, and all four margins
• 📚 Reads more than EPUB - .azw3, .mobi, .fb2, .cbz, .lit and .pdb, as long as they are DRM-free
• 🔢 Page numbers, headers and footers - optional running heads with the title, author, section and page number
• 📦 Converts a whole folder in one run - up to 100 ebooks, results zipped, and a DRM-protected book in the middle does not stop the rest
• 🔌 Runs from the API, a schedule or an integration - it is an Apify Actor, so Zapier, Make, n8n and MCP clients can call it

How to convert EPUB to PDF

1. Click Try for free and sign in to Apify.
2. Upload your .epub file in the Ebooks field, or paste a direct URL to it. You can add as many as you like.
3. Choose a Page size. A4 is the default; pick A5 if the PDF is for reading on screen rather than printing.
4. Click Start.
5. Download the PDF from the Storage tab, or from the download URL in the run's dataset record.

The remaining settings all change how the page looks rather than whether it works, so the defaults are a reasonable book. The two worth touching first are Page size and Base font size.

How to convert AZW3 to PDF

Identical steps with an .azw3 file. AZW3 (also called KF8) is Amazon's newer format and it carries real CSS, so it typesets better than old MOBI does. .azw and .prc go through the same path. DRM-protected Kindle purchases do not - see Limitations.

How to convert MOBI to PDF

Upload a .mobi file instead. MOBI is built on a cut-down HTML 3.2, so there is less structure in the source for the conversion to work with: expect simpler tables and looser layout than the same book in EPUB or AZW3. Nothing extra to configure - the Actor reads the file's own header to work out what it has.

How to convert FB2 to PDF

Upload an .fb2 file. FB2 is a single XML document rather than an archive, and it is common in Russian-language libraries. It carries clean semantic markup, so it converts tidily.

How to convert CBZ to PDF

Upload a .cbz comic archive - a zip full of page images - and each image becomes a PDF page. Turn Preserve cover aspect ratio on, which is the default, unless you want the artwork stretched to the paper. For comics, consider setting all four margins to 1 point and picking a page size close to the artwork's own proportions; A4 will otherwise letterbox a US comic page.

How to convert LIT and PDB to PDF

.lit (Microsoft Reader) and .pdb (PalmDOC) both work, as long as they are DRM-free. Both are dead formats, which is usually why someone is converting them.

What about PDF to EPUB?

That is the other direction, and this Actor does not do it. Converting a fixed-layout PDF back into reflowable text is a genuinely different and much harder problem - it means recovering reading order and structure that the PDF threw away. It will be its own Actor.

How to batch convert a folder of ebooks to PDF

Put every file in the Ebooks field - up to 100 per run - and each one is converted on its own with the same settings. This is the cheap way to do it: starting the container is most of the cost of a short run, so a hundred books in one run costs far less than a hundred runs of one book.

Each PDF is stored under its own filename, and the run also produces a ZIP of all of them, which is what you get from the Storage tab or the API response. Two books with the same filename do not overwrite each other - the second becomes book-2.pdf.

A batch does not stop on a bad file. A DRM-protected book, a dead URL or something that is not an ebook at all is recorded against that file with a reason, and the rest of the batch carries on; the run fails only if nothing at all converted. That matters for Kindle formats in particular, because most .azw and .mobi files bought from Amazon are protected, so a real folder converts partially by nature. The run's log ends with how many converted, how many failed, and how many failed specifically on DRM, and the dataset has one row per input file with an ok flag and the reason.

Batches take longer than the 300-second run-sync window, so drive them with the asynchronous Run Actor endpoint rather than run-sync.

Input

Upload your files in the Ebooks field, or pass direct URLs to them.

Field	Description
files	The ebooks to convert, as an array. A multi-file upload or a list of URLs. Up to 100.
file	A single ebook as a URL. Hidden in the Console; kept for API callers. Ignored when files is set.
paperSize	a4 (default), letter, legal, a3, a5, a6, b4, b5 or b6.
baseFontSize	Body text size in points. 0 (default) lets calibre choose to suit the page.
marginTop, marginBottom, marginLeft, marginRight	In points, 72 pt to the inch. Default 72.
pdfPageNumbers	Print a page number at the bottom of every page.
standardFont	serif (default), sans or mono for the body text.
serifFamily, sansFamily	Font family names. Must be installed in the container - see Which fonts can I use?
preserveCoverAspectRatio	On by default. Keeps the cover's proportions instead of stretching it to the page.
smartenPunctuation	Convert quotes, dashes and ellipses to typographic equivalents.
pdfHeaderTemplate, pdfFooterTemplate	HTML running heads. _PAGENUM_, _TITLE_, _AUTHOR_ and _SECTION_ are substituted.
outputFilename	Defaults to the input filename with a .pdf extension. Ignored for a batch, where every PDF keeps its own name.
bookTitle, bookAuthors	Override the PDF metadata. Separate authors with &.
fileBase64, filename	A single file as base64, for API callers who prefer JSON to a URL.
verboseLog	Include the full calibre log in the Actor log.

One quirk worth knowing, because it is calibre's and not this Actor's: a page margin of 0 means "fall back to the general margin setting", not "no margin". Use 1 if you want an edge-to-edge page. The input schema enforces a floor of 1 for that reason.

Output

Every converted PDF is written to the run's key-value store under its own filename, so it is downloadable from the Storage tab.

OUTPUT is the record the API hands back, and it holds whichever of these you actually want:

• One ebook in - OUTPUT is the PDF itself, so run-sync returns a PDF in the response body.
• Several ebooks in - OUTPUT is a ZIP of every PDF.
• A batch whose PDFs total more than 200 MB - the ZIP is skipped, because it has to be held in memory to be uploaded. OUTPUT is then a JSON manifest listing every PDF's key, size and URL. The PDFs themselves are stored exactly as before.

The dataset gets one record per input ebook, pushed as it is produced, carrying an ok flag, the source format, the source EPUB version, the page size, the page count, file sizes, the title and authors, the seconds it took, and - when ok is false - a plain-language error and a machine-readable errorKind (drm, unsupported, download, timeout, caps, budget or conversion).

How much does it cost to convert EPUB to PDF?

Job	Run time	Cost
One 10-chapter novel	13.5 s	$0.0017
10 novels in one run	62.1 s	$0.0054, or $0.0005 each

Measured 2026-08-26 on the Free plan at the default 2048 MB, read back from the run API.

A book costs a fraction of a cent, and Apify's free plan includes $5 of usage a month, so a personal library converts for nothing.

Convert a library as one run. Ten books took 62 s against 13.5 s for one, and as ten separate runs they would have cost about $0.017 rather than $0.0054 - roughly three times more. Container startup is paid once per run rather than once per book; past that, each book still costs its own few seconds of calibre time, so the saving is real but not unlimited.

Compute is most of the bill here, with storage writes and data transfer making up the rest.

Illustrated books and omnibus collections are the exception to all of the above - they are far slower than a text novel, and a very large one is better given a run of its own than dropped in the middle of a library.

How to convert EPUB to PDF with the API

run-sync returns the PDF in the response body:

curl -X POST \
  "https://api.apify.com/v2/acts/marcel-rbro~epub-to-pdf/run-sync?token=<TOKEN>" \
  -H 'Content-Type: application/json' \
  -d '{"file": "https://example.com/book.epub", "paperSize": "a4"}' \
  --output book.pdf

With the typesetting settings spelled out - A5 pages, 11 pt body text, half-inch margins and page numbers:

curl -X POST \
  "https://api.apify.com/v2/acts/marcel-rbro~epub-to-pdf/run-sync?token=<TOKEN>" \
  -H 'Content-Type: application/json' \
  -d '{
        "file": "https://example.com/book.epub",
        "paperSize": "a5",
        "baseFontSize": 11,
        "marginTop": 36, "marginBottom": 36, "marginLeft": 36, "marginRight": 36,
        "pdfPageNumbers": true
      }' \
  --output book.pdf

A batch takes a list. It will outlast the run-sync window, so start it asynchronously and collect the results afterwards:

curl -X POST \
  "https://api.apify.com/v2/acts/marcel-rbro~epub-to-pdf/runs?token=<TOKEN>" \
  -H 'Content-Type: application/json' \
  -d '{
        "files": [
          "https://example.com/one.epub",
          "https://example.com/two.azw3",
          "https://example.com/three.fb2"
        ],
        "paperSize": "a5"
      }'

If you hold the bytes rather than a URL, pass them as base64 in fileBase64. That path takes one ebook at a time:

curl -X POST \
  "https://api.apify.com/v2/acts/marcel-rbro~epub-to-pdf/run-sync?token=<TOKEN>" \
  -H 'Content-Type: application/json' \
  -d "{\"fileBase64\": \"$(base64 < book.epub)\", \"filename\": \"book.epub\"}" \
  --output book.pdf

Base64 adds about a third to the payload, so prefer a URL for large files. A URL and base64 are the two input paths: the Apify API requires Actor input to be application/json and rejects anything else with HTTP 400 before the run starts, so there is no way to post the ebook as a raw request body.

run-sync waits up to 300 seconds. Use the asynchronous Run Actor endpoint for anything larger. There are client libraries for JavaScript and Python, and the same run can be triggered on a schedule or from Zapier, Make, n8n, Slack or an MCP client.

Limitations

• Reflowable in, fixed-layout out. This is the one thing worth understanding before you start. An EPUB has no pages: the reader decides where lines and pages break, at whatever size and font the reader was set to. A PDF has nothing but pages, fixed at conversion time. So the conversion has to invent a page - a size, a font, a measure, a set of breaks - and the settings in the input are how you choose which one it invents. The text and images survive; the specific typography of the EPUB as you saw it in your reader does not, because it never existed as a fixed thing in the first place.
• Going back is much harder than coming here. A PDF made this way is a dead end for reflowing. Keep the EPUB.
• DRM-protected files fail. The Actor reads the container before converting: META-INF/rights.xml or META-INF/encryption.xml for EPUB, the encryption flag in the header for Kindle formats. The run stops with an explanation rather than producing a broken file. Most books bought from Amazon, Kobo or Google are protected. Removing that DRM is out of scope. Font obfuscation, which uses the same encryption.xml file and is not DRM, is correctly ignored.
• Complex layout comes through simplified. Multi-column pages, sidebars, drop caps and footnote placement are the usual casualties. Fixed-layout EPUBs - children's books, cookbooks, textbooks - convert least faithfully of all, because the source already had a page design and the conversion imposes a different one.
• Only fonts installed in the container are available. serifFamily and sansFamily name a system font, not a file you can upload. See the FAQ for the list.
• Up to 100 ebooks per run, within a 500 MB total input, a 200 MB ceiling on any single file, and a 55-minute whole-run conversion budget. Books the budget does not reach are recorded as failed with that reason rather than silently cut off, so raise the memory or split the batch if you hit it.
• outputFilename does not apply to a batch. Naming every output the same thing is not useful, so each PDF keeps its own filename and the option is ignored with a warning.

Other file converters

Actor	Converts
EPUB to PDF converter (this one)	EPUB, AZW3, MOBI, FB2, CBZ, LIT, PDB → PDF
MOBI to EPUB converter	MOBI, PRC, AZW, AZW3 → EPUB
Word to PDF converter	DOC, DOCX, ODT, RTF → PDF
JPG to PDF converter	JPEG, PNG → PDF
PDF to JPG converter	PDF → JPEG, PNG
HEIC to JPG converter	HEIC, HEIF, AVIF → JPEG, PNG, WebP, AVIF, TIFF
PNG to JPG converter	PNG → JPEG, WebP

FAQ

Is this EPUB to PDF converter free?

Apify's free plan includes $5 of platform usage per month and needs no card, which covers a personal library many times over. See How much does it cost for what a book actually costs.

Will the formatting survive the conversion to PDF?

The text, images, cover, links and chapter structure survive. The page does not, because the EPUB never had one - a reflowable book is re-typeset from scratch to whatever page size you pick. So the honest answer is that the result will look like a competently set book, but not like your e-reader with the settings you had it on. Fixed-layout EPUBs are the exception in the other direction: those did have a page design, and imposing a different one on them goes worse than it does for a novel.

Can I convert an EPUB I bought from Kobo, Google Play Books or Amazon?

Only if it is DRM-free, which most retail purchases are not. The Actor inspects the container and stops with an explanation rather than producing a broken file: for EPUB it looks for the Adobe ADEPT and Apple FairPlay markers and for encrypted content in META-INF/encryption.xml, and for Kindle formats it reads the encryption flag in the header. Stripping that DRM is out of scope. Project Gutenberg, Standard Ebooks, most Humble Bundles and anything from a publisher's own DRM-free store all convert fine.

My EPUB has an encryption.xml but is not DRM-protected. Does it still work?

Yes. The IDPF and Adobe font-obfuscation schemes use the same META-INF/encryption.xml file to mangle embedded fonts, and that is not DRM - calibre de-obfuscates it in passing. The Actor checks which algorithm the file names and only fails when something other than a font is actually encrypted.

Which page size should I choose?

A4 if you are going to print it outside the US, US Letter inside it. A5 or A6 if the PDF is for reading on a screen or a small e-reader - an A4 page of body text is a very wide measure to read, which is exactly why real books are not A4. A5 with a base font size around 11 pt is the closest this Actor gets to a trade paperback.

Which fonts can I use?

serifFamily and sansFamily take the name of a font installed in the container, not an upload. The container has DejaVu Serif, DejaVu Sans, DejaVu Sans Mono, Liberation Serif, Liberation Sans and Liberation Mono. Liberation Serif is metrically compatible with Times New Roman and Liberation Sans with Arial, so those are the ones to name if you are matching a house style. Leave both empty and calibre picks a sensible default. Fonts embedded in the EPUB itself are used for the text that asks for them, regardless of this setting.

How do I add page numbers, or a running header with the chapter title?

Turn on pdfPageNumbers for a plain number at the foot of each page. For anything more, use pdfHeaderTemplate and pdfFooterTemplate, which take HTML and substitute _PAGENUM_, _TITLE_, _AUTHOR_ and _SECTION_. A conventional running head is:

<div style="text-align:center; font-size:9pt">_SECTION_</div>

A footer template overrides pdfPageNumbers, so put _PAGENUM_ in the template if you use both.

Does the PDF keep the table of contents and links?

Yes. The chapter structure becomes the PDF outline - the bookmarks panel in your PDF reader - and internal and external links stay clickable. If you also want a printed contents page with page numbers in it, that is a different thing, and this Actor does not add one.

Can I convert a whole folder of ebooks at once?

Yes, that is the normal way to use it - up to 100 per run. See How to batch convert a folder of ebooks to PDF. It is also much cheaper than one run per book, because container startup is most of what a short run costs.

One book in my batch is DRM-protected. Does the whole run fail?

No. Each book is converted on its own and a failure is recorded against that book, with the reason, while the rest of the batch carries on. The run fails only if nothing converted. This is deliberate: most Kindle purchases are DRM-protected, so a folder of .azw3 files converting partially is the expected outcome rather than an error. Check the ok column in the dataset to see which ones came through.

Can an AI agent call this converter?

Yes. Every Apify Actor is exposed through the Apify MCP server, so Claude, Cursor and other MCP clients can run this conversion as a tool.

The conversion failed or the PDF looks wrong. What now?

Turn on verboseLog and re-run: the Actor log will carry calibre's own output. The common causes, roughly in order:

• A DRM-protected file. The run stops before calibre is involved and says so.
• A URL that returns an HTML page rather than the ebook - a download page or a login wall instead of the file itself.
• A truncated upload. A partly downloaded EPUB is a valid-looking zip with no central directory, and the Actor reports it as an unreadable zip rather than handing calibre a broken file.
• An omnibus or heavily illustrated book large enough to hit the 15 minute conversion timeout at low memory. Raise the memory.
• Text running off the page, which is nearly always margins set too narrow for the page size you chose.

Development

npm install
npm test      # inspector, batch bookkeeping and ZIP writer; no calibre needed
apify run     # needs calibre on your PATH

src/source.js is the inspector, src/batch.js is the batch bookkeeping - key collisions, size and time budgets, the ZIP ceiling - and src/zip.js is a small stored-entry ZIP writer shared with marcel-rbro/pdf-to-jpg. All three are pure and are what the tests cover; everything else needs the container.

The inspector reads the OCF container - the zip central directory, the mimetype entry, META-INF/container.xml and the package document - to identify the file, read its EPUB version and tell DRM apart from font obfuscation, all before calibre is involved. It also reads the PalmDB header for Kindle input.

Conversion runs calibre's ebook-convert in a Debian container. PDF output drives QtWebEngine, so the image asserts at build time that a real EPUB to PDF conversion succeeds - a Debian package that quietly loses one of its Qt libraries fails the build once rather than every run.

This Actor's code is MIT licensed; calibre is GPL-3 and runs as a separate process - see THIRD_PARTY_NOTICES.md.