CBZ → PDF is for reading a comic where no comic app exists: pages are put in the order a person counts them, and JPEG scans go into the PDF without being compressed a second time.
From real runs
| Element | Result | What happens |
|---|---|---|
| Page order | Kept | Pages are sorted by the numbers in their names, so page2 comes before page10 even when the archive stores them otherwise. A test archive stored as 10, 1, 11, 2, 12, 3 came out 1 to 12. |
| JPEG scans | Kept | All 36 JPEG pages of Terrors of the Jungle #10 were embedded byte for byte, so the PDF carries the same detail and the same compression flaws as the scans. |
| WebP pages | Changed | PDF cannot hold WebP, so WebP pages are re-encoded as JPEG at quality 90 and the result notes how many; in our mixed test archive that was two of twelve pages. |
| PNG pages | Kept | PNG pages stay lossless but are recompressed for the PDF, which makes PNG comics far slower to convert than JPEG ones. |
| Page size | Changed | Each page is sized from its pixels at 150 DPI, so a 1600 × 2357 scan becomes a 10.7 × 15.7 inch page; scans of different widths give pages of different sizes. Readers zoom to fit. |
| Double-page spreads | Kept | A spread saved as one wide image stays one wide page (936 × 648 pt in our test) rather than being cut in two. |
| ComicInfo.xml and stray files | Changed | The ComicInfo title becomes the PDF title; Thumbs.db, .DS_Store, __MACOSX folders and text notes are left out instead of turning into blank pages. |
| Damaged pages | Changed | A truncated JPEG is skipped and named in a warning, rather than appearing as a half-grey page in the PDF. |
Pages are sorted by the numbers in their names, so page2 comes before page10 even when the archive stores them otherwise. A test archive stored as 10, 1, 11, 2, 12, 3 came out 1 to 12.
All 36 JPEG pages of Terrors of the Jungle #10 were embedded byte for byte, so the PDF carries the same detail and the same compression flaws as the scans.
PDF cannot hold WebP, so WebP pages are re-encoded as JPEG at quality 90 and the result notes how many; in our mixed test archive that was two of twelve pages.
PNG pages stay lossless but are recompressed for the PDF, which makes PNG comics far slower to convert than JPEG ones.
Each page is sized from its pixels at 150 DPI, so a 1600 × 2357 scan becomes a 10.7 × 15.7 inch page; scans of different widths give pages of different sizes. Readers zoom to fit.
A spread saved as one wide image stays one wide page (936 × 648 pt in our test) rather than being cut in two.
The ComicInfo title becomes the PDF title; Thumbs.db, .DS_Store, __MACOSX folders and text notes are left out instead of turning into blank pages.
A truncated JPEG is skipped and named in a warning, rather than appearing as a half-grey page in the PDF.
We converted our Comic Book ZIP (CBZ) sample (325 KB) with Image engine on 2026-10-20. It took 59 ms and produced a 326 KB PDF file, about the same size as the input. The engine raised no warnings.
Input · sample.cbz · 325 KB
ZIP with 4 parts
Output · sample.pdf · 326 KB
PDF 1.7, 3 pages, no text layer, 3 image objects
bsdtar and img2pdf
mkdir pages && bsdtar -xf comic.cbz -C pages && find pages -type f -iname '*.jpg' -print0 | sort -zV | xargs -0 img2pdf -o comic.pdfsort -V gives number order and img2pdf embeds JPEGs without recompressing. Add PNG patterns if needed, convert WebP pages first, and extract untrusted archives in an empty folder.
MuPDF (mutool)
mutool convert -o comic.pdf comic.cbzQuick for clean archives, but mutool 1.21 gave up on our mixed archive with 'unknown image file format' because of its WebP pages.
The HTTP API and the MCP server run the exact engine this page uses (Image engine); the response carries the same warnings the web converter shows. Check support first with GET /api/v1/pairs?from=cbz&to=pdf.
curl -X POST https://filetypeconverters.com/api/v1/convert \
-H "Authorization: Bearer $FTC_API_KEY" \
-F "file=@input.cbz" -F "from=cbz" -F "to=pdf" \
-F 'options={"width":0}' \
--output output.pdf| Key | Applies to | Type | Default |
|---|---|---|---|
| width | Writing PDF | number | 0 |
| height | Writing PDF | number | 0 |
| fit | Writing PDF | contain · cover · fill | contain |
| quality | Writing PDF | number | 90 |
| background | Writing PDF | string | #ffffff |
| stripMetadata | Writing PDF | boolean | true |
They sort file names letter by letter, so 1, 10, 11, 2 comes out in that order. We compare the numbers inside names as numbers, which fixes unpadded names like page2.jpg and page10.jpg without renaming anything.
Not with JPEG pages: they are copied in unchanged, so the PDF is about the size of the page images themselves, which came to 3.6 percent more than the CBZ for Terrors of the Jungle #10. Set width to resample pages if you need a smaller file.
No. A spread saved as one wide image becomes one wide PDF page, which is how most PDF readers show it best in landscape.