c #. System.Windows.Forms.WebBrowser and the text code for the IFrame

The DocumentText property returns what you are asking for - the document code. The code of the content of the iframes is not included in it, since the iframe is more than just a mechanism for substituting one HTML into another (*). To get a document along with the contents of iframes, the only way is to go through all the iframes, take their contents (for example, through the MSHTML.IHTMLIFrameElement3 interface) and substitute it into the original document. Clearly, you can not get the content if the iframe looks to a different domain.

Something like this:

 using System; using System.Collections.Generic; using System.ComponentModel; using System.Runtime.InteropServices; using System.Text; using System.Windows.Forms; namespace WebBrowserTest { public partial class Form1 : Form { public Form1() { InitializeComponent(); } public String DomDocumentText { get { var document = webBrowser1.Document; string returnstr = ""; MSHTML.IHTMLDocument3 doc3 = null; MSHTML.HTMLDocument new_doc = null; MSHTML.IHTMLDocument2 doc2 = null; MSHTML.IHTMLElementCollection elems = null; MSHTML.IHTMLDocument3 new_doc3 = null; MSHTML.IHTMLElementCollection elems_new = null; MSHTML.IHTMLDocument3 child_doc = null; MSHTML.IHTMLElement content = null; MSHTML.IHTMLElement content_new = null; MSHTML.IHTMLElementCollection elem_col = null; try { doc3 = (MSHTML.IHTMLDocument3)document.DomDocument;//исходный документ new_doc = new MSHTML.HTMLDocument();//копия документа doc2 = new_doc as MSHTML.IHTMLDocument2; doc2.write(webBrowser1.DocumentText);//создаем копию документа //получаем все iframe в документе и его копии... elems = doc3.getElementsByTagName("iframe"); new_doc3 = new_doc as MSHTML.IHTMLDocument3; elems_new = (new_doc3).getElementsByTagName("iframe"); int i = 0; foreach (MSHTML.IHTMLIFrameElement3 elem in elems) //для каждого iframe... { try { child_doc = elem.contentDocument as MSHTML.IHTMLDocument3; elem_col = child_doc.getElementsByTagName("body"); if (elem_col == null || elem_col.length == 0) { i++; continue; } content = (MSHTML.IHTMLElement)elem_col.item(0); string str = (content).innerHTML;//получаем содержимое iframe content_new = elems_new.item(i) as MSHTML.IHTMLElement; (content_new).outerHTML = str;//заменяем iframe на его содержимое i++; } catch (Exception ex) { //для iframe, URL которых в другом домене, будет исключение //HRESULT: 0x80070005 (E_ACCESSDENIED) System.Diagnostics.Debug.WriteLine("Can't process iframe " + i.ToString()); System.Diagnostics.Debug.WriteLine(ex.Message); } finally { //Очистка ресурсов, задействованных в цикле... if (child_doc != null) { Marshal.ReleaseComObject(child_doc); child_doc = null; } if (elem_col != null) { Marshal.ReleaseComObject(elem_col); elem_col = null; } if (content != null) { Marshal.ReleaseComObject(content); content = null; } if (content_new != null) { Marshal.ReleaseComObject(content_new); content_new = null; } } }//end foreach returnstr = new_doc.documentElement.innerHTML; return returnstr; } finally { //Окончательная очистка ресурсов... if (doc3 != null) Marshal.ReleaseComObject(doc3); if (new_doc != null) Marshal.ReleaseComObject(new_doc); if (doc2 != null) Marshal.ReleaseComObject(doc2); if (elems != null) Marshal.ReleaseComObject(elems); if (new_doc3 != null) Marshal.ReleaseComObject(new_doc3); if (elems_new != null) Marshal.ReleaseComObject(elems_new); if (child_doc != null) Marshal.ReleaseComObject(child_doc); if (content != null) Marshal.ReleaseComObject(content); if (content_new != null) Marshal.ReleaseComObject(content_new); if (elem_col != null) Marshal.ReleaseComObject(elem_col); } } } private void button1_Click(object sender, EventArgs e) { MessageBox.Show(DomDocumentText); } } } 

To use, you need to connect the library of IE COM objects (in Visual Studio Add a link -> COM -> Microsoft HTML Object Library , or manually specify the file mshtml.tlb).

Notes

(*) - For the iframe, the browser creates a separate window inside the main window, which can, in general, display not HTML, but a document of another type. Therefore, in the DOM model, the InnerHTML property of the iframe does not correspond to its content, but to the replacement text for browsers without iframe support, which can be placed inside the iframe tag (which is rarely used in our time).

To solve my problem, I used JavaScript indicated in the question:

 documentHtml = webBrowser.DomDocumentText; // extract frames content (by frame id) var framesAsHtmlElements = webBrowser.Document.GetElementsByTagName("iframe"); foreach(HtmlElement frame in framesAsHtmlElements) { String id = frame.Id; if (String.IsNullOrEmpty(id) == false) { string jCode = "var iframe = document.getElementById('" + id + "'); var innerDoc = iframe.contentDocument || iframe.contentWindow.document; innerDoc.documentElement.innerHTML"; String frameHtml = webBrowser.WebBrowser.Document.InvokeScript("eval", new object[] { jCode }) as String; if (String.IsNullOrEmpty(frameHtml) == false) { // ((?im)<iframe(([.]|[^/]|\r\n|\r|\n)*?id=["']frame-id["'](.|\r\n|\r|\n)*?)(/>|</iframe>)+?) Regex regEx = new Regex(String.Format(@"(<iframe(([.]|[^/]|\r\n|\r|\n)*?id=[""']{0}[""'](.|\r\n|\r|\n)*?)(/>|</iframe>)+?)", id), RegexOptions.Multiline | RegexOptions.IgnoreCase ); Match match = regEx.Match(documentHtml); if (match.Success) { int pos = match.Index; int length = match.Length; documentHtml = documentHtml.Substring(0, pos) // before <iframe> + "<iframe" + match.Groups[2].Value + "/>" + frameHtml + "</iframe>" + documentHtml.Substring(pos + length); ; } } } }