My web app needs to catch every non‑existent path as a search string, because I’m implementing an HttpModule that handles URLs in an SEO‑friendly way, e.g.:

http://contoso.com/search-string

But the Microsoft AJAX ScriptModule interfered, causing runtime errors like “AJAX Framework failed to load…”. The fix was to register my module after the ScriptModule and selectively bypass AJAX requests.

<add name="ScriptModule"
     type="System.Web.Handlers.ScriptModule, System.Web.Extensions, ..." />
<add name="MyModule" type="MyModule" />

using System;
using System.Linq;
using System.Web;

public class MyModule : IHttpModule
{
  public void Init(HttpApplication context)
  {
    context.PostResolveRequestCache += Application_OnAfterProcess;
  }

  private void Application_OnAfterProcess(object source, EventArgs e)
  {
    var application = (HttpApplication)source;
    var context = application.Context;

    if (context.Request.Headers["x-microsoftajax"] == null)
    {
      if (!System.IO.File.Exists(application.Request.PhysicalPath) &&
          !application.Request.Url.ToString().Contains(".axd") &&
          !application.Request.Url.ToString().Contains(".asmx"))
      {
        string newUrl = "~/Search.aspx?q=" +
          context.Server.UrlEncode(application.Request.Url.Segments.Last());
        // other logic here…
        context.RewritePath(newUrl);
      }
    }
  }

  public void Dispose() { }
}

This runs at PostResolveRequestCache and, when the special x-microsoftajax header is absent, it rewrites unknown paths to your search page while letting AJAX handlers (*.axd, *.asmx) pass through normally.