Scroll to navigation

libcurl(3) libcurl URL interface libcurl(3)


libcurl-url - URL interface overview


The URL interface provides functions for parsing and generating URLs.


You still only include <curl/curl.h> in your code.


Create a handle that holds URL info and resources with curl_url(3):

CURLU *h = curl_url();


When done with it, clean it up with curl_url_cleanup(3):



When you need a copy of a handle, just duplicate it with curl_url_dup(3):

CURLU *nh = curl_url_dup(h);


By "setting" a URL to the handle with curl_url_set(3), the URL is parsed and stored in the handle. If the URL is not syntactically correct it will return an error instead.

rc = curl_url_set(h, CURLUPART_URL,
"", 0);

The zero in the fourth argument is a bitmask for changing specific features.

If successful, this stores the URL in its individual parts within the handle.


When a handle already contains info about a URL, setting a relative URL will make it "redirect" to adapt to it.

rc = curl_url_set(h, CURLUPART_URL, "../test?another", 0);


The `CURLU` handle represents a URL and you can easily extract that with curl_url_get(3):

char *url;
rc = curl_url_get(h, CURLUPART_URL, &url, 0);

The zero in the fourth argument is a bitmask for changing specific features.


When a URL has been parsed or parts have been set, you can extract those pieces from the handle at any time.

rc = curl_url_get(h, CURLUPART_HOST, &host, 0);
rc = curl_url_get(h, CURLUPART_SCHEME, &scheme, 0);
rc = curl_url_get(h, CURLUPART_USER, &user, 0);
rc = curl_url_get(h, CURLUPART_PASSWORD, &password, 0);
rc = curl_url_get(h, CURLUPART_PORT, &port, 0);
rc = curl_url_get(h, CURLUPART_PATH, &path, 0);
rc = curl_url_get(h, CURLUPART_QUERY, &query, 0);
rc = curl_url_get(h, CURLUPART_FRAGMENT, &fragment, 0);

Extracted parts are not URL decoded unless the user also asks for it with the CURLU_URLDECODE flag set in the fourth bitmask argument.

Remember to free the returned string with curl_free(3) when you are done with it!


A user set individual URL parts, either after having parsed a full URL or instead of parsing such.

rc = curl_url_set(urlp, CURLUPART_HOST, "", 0);
rc = curl_url_set(urlp, CURLUPART_SCHEME, "https", 0);
rc = curl_url_set(urlp, CURLUPART_USER, "john", 0);
rc = curl_url_set(urlp, CURLUPART_PASSWORD, "doe", 0);
rc = curl_url_set(urlp, CURLUPART_PORT, "443", 0);
rc = curl_url_set(urlp, CURLUPART_PATH, "/index.html", 0);
rc = curl_url_set(urlp, CURLUPART_QUERY, "name=john", 0);
rc = curl_url_set(urlp, CURLUPART_FRAGMENT, "anchor", 0);

Set parts are not URL encoded unless the user asks for it with the `CURLU_URLENCODE` flag.


An application can append a string to the right end of the query part with the `CURLU_APPENDQUERY` flag to curl_url_set(3).

Imagine a handle that holds the URL ``. An application can then add the string `hat=1` to the query part like this:

rc = curl_url_set(urlp, CURLUPART_QUERY, "hat=1", CURLU_APPENDQUERY);

It will even notice the lack of an ampersand (`&`) separator so it will inject one too, and the handle's full URL will then equal ``.

The appended string can of course also get URL encoded on add, and if asked to URL encode, the encoding process will skip the '=' character. For example, append `candy=N&N` to what we already have, and URL encode it to deal with the ampersand in the data:

rc = curl_url_set(urlp, CURLUPART_QUERY, "candy=N&N",

Now the URL looks like`


The URL API was introduced in libcurl 7.62.0.


curl_url(3), curl_url_cleanup(3), curl_url_get(3), curl_url_dup(3), curl_url_set(3), curl_url_strerror(3), CURLOPT_URL(3)

May 17, 2022 libcurl 7.85.0