wget Rename File: How to Save a Download Under a Different Name (wget -O)

Updated: 4 November 2026 · Applies to: GNU Wget 1.25.0 (Debian 13, tested), Wget2 2.2.1 (Fedora 44, tested) and curl 8.14.1; Ubuntu 24.04 (Wget 1.21.4), AlmaLinux 9 (1.21.1) and AlmaLinux 10 (1.24.5) ship GNU Wget with the same options

To rename a file while downloading it with wget, use -O (capital letter O, long form --output-document) followed by the new name:

wget -O ca-bundle.pem https://curl.se/ca/cacert.pem

wget downloads cacert.pem but saves it as ca-bundle.pem:

HTTP request sent, awaiting response... 200 OK
Length: 188900 (184K) [application/x-pem-file]
Saving to: 'ca-bundle.pem'
...
'ca-bundle.pem' saved [188900/188900]

Put the URL in quotes if it contains ? or &, so the shell does not interpret those characters.

Rename and save in another folder

Give -O a full path:

wget -O /root/downloads/ca-bundle.pem https://curl.se/ca/cacert.pem

• The folder must already exist. Otherwise wget stops with nodir/ca.pem: No such file or directory (create it first with mkdir -p).
• -P (directory prefix) has no effect together with -O: wget -P /root/downloads -O ca-bundle.pem URL saved ca-bundle.pem in the current folder in our test, not in /root/downloads. Use -P alone when you want to keep the original name in another folder (see download files to a specific directory with wget).

Resume a renamed download

-c (continue) works with -O, as long as you use the same name again. We cut a 2.7 MB file at 1 MB and resumed it:

$ wget -c -O curl.tar.xz https://curl.se/download/curl-8.14.1.tar.xz
HTTP request sent, awaiting response... 206 Partial Content
Length: 2817248 (2.7M), 1817248 (1.7M) remaining [application/x-xz]
Saving to: 'curl.tar.xz'
        [ skipping 950K ]
...
'curl.tar.xz' saved [2817248/2817248]

The SHA256 checksum matched the complete download. Resuming needs a server that supports HTTP range requests.

What -O really does

The wget manual says -O is not meant as "use this name instead of the one in the URL": it works like shell redirection, so wget -O file URL behaves like wget -O - URL > file. In practice:

• An existing file is overwritten without a question (a file with the same name was replaced in our test).
• A failed download leaves an empty file. A URL that returned an error (exit code 8) still left a 0-byte missing.pem. Check the exit code (echo $?, 0 means success) in scripts.
• Several URLs with one -O are joined into one file. wget -O both.txt URL1 URL2 wrote both downloads into both.txt. Use one wget command per file instead.
• With -nc (no clobber), wget refuses if the file exists: File 'ca-bundle.pem' already there; not retrieving.
• -O - sends the download to standard output, for example wget -qO - URL | tar xzf - (see download and extract tar files with one command).
• According to the wget manual, -N (timestamp checking) is not supported with -O, and with -r or -p all content goes into that single file.

Let the server choose the name

Some download links do not end in a useful file name. Without -O, wget names the file after the last part of the URL you typed, query string included: wget "https://www.gnu.org/licenses/gpl-3.0.txt?ref=kb" saved gpl-3.0.txt?ref=kb. Two options change this:

• --content-disposition uses the file name the server sends in the Content-Disposition header.
• --trust-server-names names the file after the last part of the URL wget was redirected to.

For the GitHub link https://github.com/jqlang/jq/archive/refs/tags/jq-1.7.1.tar.gz (which redirects and sends content-disposition: attachment; filename=jq-jq-1.7.1.tar.gz) we got three different names:

wget URL                          -> jq-1.7.1.tar.gz
wget --content-disposition URL    -> jq-jq-1.7.1.tar.gz
wget --trust-server-names URL     -> jq-1.7.1

When you need a predictable name, -O is still the safest choice.

File already exists: why wget adds .1

Without -O, downloading the same file twice does not overwrite it. wget saves the second copy as cacert.pem.1, then .2 and so on. Use -O name to overwrite, -nc to skip existing files, or -c to resume a partial file.

Rename several downloads at once

Put each URL and its new name on one line in list.txt:

https://curl.se/ca/cacert.pem ca-bundle.pem
https://curl.se/ca/cacert.pem.sha256 ca-bundle.sha256

Then loop over it:

while read -r url name; do wget -q -O "$name" "$url"; done < list.txt

Both files were saved under the new names. (wget -i list.txt downloads a list too, but keeps the original names.)

The same with curl and Wget2

• curl: curl -L -o newname URL saves under your name (-L follows redirects). curl -LO URL keeps the URL name and curl -LOJ URL uses the Content-Disposition name (it gave jq-jq-1.7.1.tar.gz for the GitHub link above). More in tips for the curl command.
• Fedora: the wget command on Fedora 44 is Wget2 (package wget2-wget, GNU Wget2 2.2.1). -O, -c and --content-disposition gave the same results; its messages look different (Saving 'ca-bundle.pem'), and resuming a complete file prints The file is already fully retrieved; nothing to do.

You can also rename after the download with mv cacert.pem ca-bundle.pem. To keep big downloads from using all your bandwidth, see limit download speed with wget and curl.

Official documentation: wget(1) manual and curl manual.