The code below is provided under an MIT license, but if you do make any updates it would be great if you could send them back.

Usage:

   //Example to download google's source and print out the urls of all the images
   NSError * error = nil;
   HTMLParser * parser = [[HTMLParser alloc] initWithContentsOfURL:[NSURL URLWithString:@"http://www.google.com"] error:&error];

   if (error) {
     NSLog(@"Error: %@", error);
     return;
  }
  HTMLNode * bodyNode = [parser body]; //Find the body tag

  NSArray * imageNodes = [bodyNode findChildTags:@"img"]; //Get all the 

 for (HTMLNode * imageNode in imageNodes) { //Loop through all the tags
       NSLog(@"Found image with src: %@", [imageNode getAttributeNamed:@"src"]); //Echo the src=""
  }

  [parser release];

You can grab a copy of the code here on github:
http://github.com/zootreeves/Objective-C-HMTL-Parser
git@github.com:zootreeves/Objective-C-HMTL-Parser.git

Note:

-(NSString*)rawContents; does not work, you need to use NSString * rawContentsOfNode(xmlNode * node, htmlDocPtr doc); to dump the entire html contents of a node.