Martin Thoma A blog about Code, Politics, and my life

Get PDF pages

· by

Once in a while, I want to get the total number of PDF pages of a document. You can do that with the command

$ pdfinfo document.pdf | grep Pages | awk -F':' '{gsub(/ /, "", $0);print $2}'

It works like this:

  1. Call pdfinfo to get a lot of information, including a line which begins with the string "Pages".
  2. Get only the line with the string "Pages"
  3. Split that line at ":", remove all white space and print only that after the first ":" (and before a second ":")

Now we know how to get the number of pages of a PDF document for a single document. But what if you want to get it for all documents within a folder?

$ for i in *.pdf; do pdfinfo $i | grep Pages | awk -F':' '{gsub(/ /, "", $0);print $2}'; done | paste -sd+ - | bc

It works like this:

  1. for i in *.pdf; do ...; done goes through all files ending with .pdf and prints the number of pages of the single documents.
  2. paste -sd+ - makes sure that the single lines have a + in between
  3. bc calculates the expression which is now a string like 18+42+9

Update 2026: pdfly

Nowadays I use pdfly, a command line tool written in pure Python on top of pypdf. You can install it with pip install pdfly. Its meta command shows the number of pages together with other metadata:

$ pdfly meta document.pdf
                     Operating System Data
┏━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃         Attribute ┃ Value                                   ┃
┡━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│         File Name │ /home/moose/Documents/Mega/document.pdf │
│  File Permissions │ -rw-rw-r--                              │
│         File Size │ 427,286 bytes                           │
│     Creation Time │ 2026-09-13 21:54:16                     │
│ Modification Time │ 2026-09-13 21:53:42                     │
│       Access Time │ 2026-09-26 08:15:15                     │
└───────────────────┴─────────────────────────────────────────┘
                                    PDF Data
┏━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃          Attribute ┃ Value                                                   ┃
┡━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│            Creator │ Simple Scan 46.0                                        │
│              Pages │ 1                                                       │
│          Encrypted │ None                                                    │
│   PDF File Version │ %PDF-1.3                                                │
│        Page Layout │                                                         │
│          Page Mode │                                                         │
│             PDF ID │ ID1=b'm\xc2\xbf\xe2\x80\xba_>|I\xe2\x80\x9a.\xc3\xbe\x… │
│                    │ ID2=b'm\xc2\xbf\xe2\x80\xba_>|I\xe2\x80\x9a.\xc3\xbe\x… │
│ Fonts (unembedded) │                                                         │
│   Fonts (embedded) │                                                         │
│        Attachments │ []                                                      │
│             Images │ 1 images (425,148 bytes)                                │
└────────────────────┴─────────────────────────────────────────────────────────┘

For scripts, pdfly can print the metadata as JSON, so jq can pick the page count:

$ pdfly meta --output json document.pdf | jq .pages
1

The page count of all documents within a folder:

$ for i in *.pdf; do pdfly meta --output json "$i"; done | jq -s 'map(.pages) | add'

It works like this:

  1. The loop prints one JSON object per document.
  2. jq -s (slurp) reads all of them into one array.
  3. map(.pages) takes the page count of each document and add sums them up.