How to save AWS CLI output to a file

Saving AWS CLI output to a file keeps API responses, inventory snapshots, and one-off checks available after the terminal session ends. A saved file is easier to attach to a ticket, compare across repeated runs, or hand to another tool than copied terminal text.

Shell redirection writes stdout to disk, while --output, --query, and --no-cli-pager decide the exact format, fields, and terminal behavior before the file is created. That makes it possible to save a full json response, a narrowed text value, or the same command result to both the screen and a file.

The saved file receives only stdout. Current AWS documentation recommends pairing --output text with --query for predictable field ordering, and AWS CLI errors stay on stderr unless they are redirected separately. The commands below use ec2 describe-regions --generate-cli-skeleton output so the save workflow can be demonstrated locally without live credentials, but generated skeleton output is not stable across CLI versions.

1. Save the command output as json when the file should keep the normal structured response body.

   $ aws ec2 describe-regions --generate-cli-skeleton output --output json --no-cli-pager > regions.json

   --output json overrides any saved profile formatter for that one command, and --no-cli-pager keeps the response from opening in the default pager before the shell writes the file.

2. Open the saved file and confirm the response body was written.

   $ cat regions.json
   {
       "Regions": [
           {
               "OptInStatus": "OptInStatus",
               "Geography": [
                   {
                       "Name": "Name"
                   }
               ],
               "RegionName": "RegionName",
               "Endpoint": "Endpoint"
           }
       ]
   }

3. Save only the field list that matters when the next step needs a smaller text file instead of the full response.

   $ aws ec2 describe-regions --generate-cli-skeleton output --query 'Regions[].RegionName' --output text --no-cli-pager > region-names.txt

   Current AWS documentation recommends using --query with --output text so the saved columns and values stay predictable.

4. Use tee when the same command result should stay visible in the terminal and be written to a file at the same time.

   $ aws ec2 describe-regions --generate-cli-skeleton output --query 'Regions[].RegionName' --output text --no-cli-pager | tee region-names-live.txt
   RegionName

   tee is useful during manual checks because the terminal view and the saved file come from the same command run.

5. Redirect stderr to a separate log when the response body and the diagnostic trace must be kept in different files.

   $ aws ec2 describe-regions --generate-cli-skeleton output --output json --debug --no-cli-pager > regions-debug.json 2> regions-debug.log

   Replace > with >> when the response file should grow across repeated runs instead of being replaced.

6. Check that the response file and the debug log were written to separate paths.

   $ ls -l regions-debug.json regions-debug.log
   -rw-r--r--  1 user  staff    277 Apr 19 20:56 regions-debug.json
   -rw-r--r--  1 user  staff  16619 Apr 19 20:56 regions-debug.log

   regions-debug.json contains stdout, while regions-debug.log contains the verbose debug trace from stderr.