How to save cURL output to a file

Saving cURL output to local files keeps response bodies, download artifacts, and API samples available after terminal scrollback is gone. Saved copies can be inspected later, handed to another tool, or attached to a ticket without repeating the same request.

By default, cURL writes the response body to standard output. --output writes to a chosen filename, --remote-name uses the last path segment from the URL, and --write-out with filename_effective prints the filename that cURL actually wrote.

Local files are overwritten by default when the same target name is reused. --no-clobber and --remove-on-error reduce that risk, but both options require cURL 7.83.0 or later and older systems may reject them as unknown options.

1. Save the response body to a chosen filename with --output.

   $ curl --silent --output message.txt --write-out '%{filename_effective}\n' https://httpbin.org/base64/SFRUUEJJTiBpcyBhd2Vzb21l
   message.txt

   filename_effective prints the path that cURL wrote after the transfer finishes.

2. Read the saved file to confirm that the response body went to disk instead of the terminal.

   $ cat message.txt
   HTTPBIN is awesome

   Reading the saved file immediately catches cases where the target path or the returned content is not the data expected.

3. Save with the URL's filename when the last path segment is already the local name wanted.

   $ curl --silent --remote-name --write-out '%{filename_effective}\n' https://httpbin.org/robots.txt
   robots.txt

   --remote-name uses the last path segment from the URL, so this request creates robots.txt in the current directory unless --output-dir is also used.

4. Prevent an existing file from being overwritten with --no-clobber.

   $ curl --silent --remote-name --no-clobber --write-out '%{filename_effective}\n' https://httpbin.org/robots.txt
   robots.txt.1

   --no-clobber appends a numeric suffix such as .1 when the original name already exists, and it requires cURL 7.83.0 or later.

5. Remove failed output automatically with --fail and --remove-on-error.

   $ curl --fail --silent --show-error --remove-on-error --output missing.txt https://httpbin.org/status/404
   curl: (56) The requested URL returned error: 404

   --fail turns an HTTP error response into a curl failure, and --remove-on-error deletes the incomplete output file instead of leaving stale content behind. The exact curl error code can vary by build or server response.

6. Check that the failed transfer did not leave the target file behind.

   $ ls missing.txt
   ls: missing.txt: No such file or directory

   If cURL reports unknown option –remove-on-error, the local build is older than 7.83.0 and failed saves need manual cleanup.