[c#] How can I get a web site's favicon?

Simple enough question: I've created a small app that is basically just a favourites that sits in my system tray so that I can open often-used sites/folders/files from the same place. Getting the default icons from my system for known file types isn't terribly complicated, but I don't know how to get the favicon from a website. (SO has the grey->orange stack icon in the address bar for instance)

Does anyone know how I might go about that?

This question is related to c# .net-4.0 favicon

The answer is


You can do it without programming. Just open the web site, right-click and select "view source" to open the HTML code of that site. Then in the text editor search for "favicon" - it will direct you to something looking like

<link rel="icon" href='/SOMERELATIVEPATH/favicon.ico' type="image/x-icon" />

take the string in href and append it to the web site's base URL (let's assume it is "http://WEBSITE/"), so it looks like

http://WEBSITE/SOMERELATIVEPATH/favicon.ico

which is the absolute path to the favicon. If you didn't find it this way, it can be as well in the root in which case the URL is http://WEBSITE/favicon.ico.

Take the URL you determined and insert it into the following code:

<html>
  <head>
   <title>Capture Favicon</title>   
  </head>
  <body>
    <a href='http://WEBSITE/SOMERELATIVEPATH/favicon.ico' alt="Favicon"/>Favicon</a> 
  </body>
</html>

Save this HTML code locally (e.g. on your desktop) as GetFavicon.html and then double-click on it to open it. It will display only a link named Favicon. Right-click on this link and select "Save target as..." to save the Favicon on your local PC - and you're done!


In 2020, using duckduckgo.com's service from the CLI

curl -v https://icons.duckduckgo.com/ip2/<website>.ico > favicon.ico

Example

curl -v https://icons.duckduckgo.com/ip2/www.cdc.gov.ico > favicon.ico

http://realfavicongenerator.net/favicon_checker?site=http://stackoverflow.com gives you favicon analysis stating which favicons are present in what size. You can process the page information to see which is the best quality favicon, and append it's filename to the URL to get it.


It's a good practice to minimize the number of requests each page needs. So if you need several icons, yandex can do a sprite of favicons in one query. Here is an example http://favicon.yandex.net/favicon/google.com/stackoverflow.com/yandex.net/


You can get the favicon URL from the website's HTML.

Here is the favicon element:

<link rel="icon" type="image/png" href="/someimage.png" />

You should use a regular expression here. If no tag found, look for favicon.ico in the site root directory. If nothing found, the site does not have a favicon.



Using jquery

var favicon = $("link[rel='shortcut icon']").attr("href") ||
              $("link[rel='icon']").attr("href") || "";

You can use Getfv.co :

To retrieve a favicon you can hotlink it at... http://g.etfv.co/[URL]

Example for this page : http://g.etfv.co/https://stackoverflow.com/questions/5119041/how-can-i-get-a-web-sites-favicon

Download content and let's go !

Edit :

Getfv.co and fvicon.com look dead. If you want I found a non free alternative : grabicon.com.


HttpWebRequest w = (HttpWebRequest)HttpWebRequest.Create("http://stackoverflow.com/favicon.ico");

w.AllowAutoRedirect = true;

HttpWebResponse r = (HttpWebResponse)w.GetResponse();

System.Drawing.Image ico;
using (Stream s = r.GetResponseStream())
{
    ico = System.Drawing.Image.FromStream(s);
}

ico.Save("favicon.ico");

The first thing to look for is /favicon.ico in the site root; something like WebClient.DownloadFile() should do fine. However, you can also set the icon in metadata - for SO this is:

<link rel="shortcut icon"
   href="http://sstatic.net/stackoverflow/img/favicon.ico">

and note that alternative icons might be available; the "touch" one tends to be bigger and higher res, for example:

<link rel="apple-touch-icon"
   href="http://sstatic.net/stackoverflow/img/apple-touch-icon.png">

so you would parse that in either the HTML Agility Pack or XmlDocument (if xhtml) and use WebClient.DownloadFile()

Here's some code I've used to obtain this via the agility pack:

var favicon = "/favicon.ico";
var el=root.SelectSingleNode("/html/head/link[@rel='shortcut icon' and @href]");
if (el != null) favicon = el.Attributes["href"].Value;

Note the icon is theirs, not yours.


The SHGetFileInfo (Check pinvoke.net for the signature) lets you retrieve a small or large icon, just as if you were dealing with a file/folder/Shell item.


Sometimes we can't get the favicon image with the purposed solution as some websites use .png or other image extensions. Here is the working solution.

  1. Open your website with a firefox browser.
  2. Right-click on the website and click the "View page info" option from the list.
  3. It will open up a dialog and click on the "Media" tab.
  4. In that tab you will see all the images including favicon.
  5. Select the favicon.ico image or click through the images to see which image is used as favicon. Some websites use .png images as well.
  6. Then click on the "Save As" button and you should be good to go.

thanks!


This is a late answer, but for completeness: it is difficult to get even close to fetching 90% all favicons.

A while ago I wrote a WordPress plugin which attempts to get closer to 100%.

This is how it works:

  1. It starts by searching existing favicon repositories such as Google favicons and GetFavicons for the favicon.

  2. If none of them returns an icon, the plugin attempts to get the icon itself. This involves traversing several pages on the domain.

  3. The plugin then inspects the physical image file, because on some servers files get returned with the incorrect mime types.

The code is still not perfect because in the details you will find many weird situations: people have wrongly coded paths, e.g. img/favicon.ico where img is not in the root, duplicate headers in HTML output, different server responses from the head and body etc.

The core of the fetching part is here so you can reverse-engineer it, but be aware that validating the response should be done (checking image filetype, mime etc.).


Updated 2020

Here are three services you can use in 2020 onwards

<img height="16" width="16" src='https://icons.duckduckgo.com/ip3/www.google.com.ico' />

<img height="16" width="16" src='http://www.google.com/s2/favicons?domain=www.google.com' />

<img height="16" width="16" src='https://api.statvoo.com/favicon/?url=google.com' />