Documentation about how to setup your own custom controllers and routes that need to exist alongside the Umbraco pipeline

Where to put your routing logic?

There's two places you can specify your routing, depending on whether it's in the context of a package, or your own site. If it's your own site you can do it in the Program.cs file, within the WithEndpoints method call like so:

app.UseUmbraco()
    .WithMiddleware(u =>
    {
        u.UseBackOffice();
        u.UseWebsite();
    })
    .WithEndpoints(u =>
    {
        // This is where to put the custom routing

        u.UseBackOfficeEndpoints();
        u.UseWebsiteEndpoints();
    });

If you're creating a package you won't have access to the Program.cs file. Instead you must use a composer, for an example of this, see the example below.

User defined routes

Umbraco doesn't interfere with any user defined routes that you wish to have. Your custom routes to your own custom controllers will work perfectly and seamlessly alongside Umbraco's routes.

Custom routes within the Umbraco pipeline

For a request to be considered executing in an Umbraco context, and therefore the Umbraco pipeline, it needs to have an HTTP request feature with the type UmbracoRouteValues, all the information required for Umbraco to handle the request is stored there. The question is now, how do we add this request feature? There's three possibilities:

1. Do it manually - This requires that you have a custom route, controller, even middleware, and manually assign the UmbracoRouteValues as an HTTP request feature, however you see fit. To create an UmbracoRouteValues object generally requires: IUmbracoContextAccessor (to access the CleanedUmbracoUrl), IPublishedRouter (to create the IPublishedRequestBuilder), IPublishedRequestBuilder (to set the published content and to build the IPublishedRequest), IPublishedRequest to assign to the UmbracoRouteValues. As you can see this is a lot of work, but luckily there's some much easier ways.

2. Route a custom controller that implements the IVirtualPageController interface, assigning the UmbracoRouteValues to the HTTP requests will then be taken care of for you.

3. Route a custom controller with conventional routing, using the typical call to endpoints.MapControllerRoute, and then call .ForUmbracoPage() with an action for finding content on what MapControllerRoute returns, now UmbracoRouteValues will automatically be applied to any request to that controller.

Don't fret if this all seems a bit overwhelming, we'll be going through an example of the last two options.

Custom route with IVirtualPageController

As mentioned, with this approach we need to implement the IVirtualPageController interface, this interface only has one method FindContent which accepts an ActionExecutingContext:

IPublishedContent FindContent(ActionExecutingContext actionExecutingContext);

It can also be helpful to inherit from the UmbracoPageController since this includes some useful helper methods such as CurrentPage, do however note that it is not possible to inherit from RenderController when doing custom routes like this.

Let's create a shop controller, with an Index action showing all our products. We will also add a Product action showing custom data about the product that could exists outside Umbraco. A common approach in a scenario like this is to have a "real" Umbraco node as a starting point. In this example we will use an empty "Products" Document Type with a Collection, and "Product" Document Type containing a Stock-Keeping Unit (SKU). We also need some content based on those document types, a "Products" content node, which contains two product nodes, each with their own SKU.

After that bit of setup we can go ahead and create our shop controller which inherits from UmbracoPageController and implements IVirtualPageController, it'll look like this:

using Microsoft.AspNetCore.Mvc.Filters;
using Microsoft.AspNetCore.Mvc.ViewEngines;
using Microsoft.Extensions.Logging;
using Umbraco.Cms.Core.Models.PublishedContent;
using Umbraco.Cms.Web.Common.Controllers;

namespace RoutingDocs.Controllers;

public class ShopController : UmbracoPageController, IVirtualPageController
{
    public ShopController(
        ILogger<UmbracoPageController> logger,
        ICompositeViewEngine compositeViewEngine)
        : base(logger, compositeViewEngine)
    { }

    public IPublishedContent FindContent(ActionExecutingContext actionExecutingContext)
    { }
}

Now you'll see that FindContent is complaining because we're not returning anything yet, but let's start by creating our to action methods that FindContent will find content for.

First off we have the Index method:

[HttpGet]
public IActionResult Index()
{
    // CurrentPage (IPublishedContent) will be the content returned
    // from the FindContent method.

    // return the view with the IPublishedContent
    return View(CurrentPage);
}

With this method we return the view with the content found by the FindContent method. This can then be used to list all the children in the view with Model.Children.

Next we have our Product method:

[HttpGet]
public IActionResult Product(string id)
{
    // CurrentPage (IPublishedContent) will be the content returned
    // from the FindContent method.

    // One example of using a custom route would be to include additional
    // model information based on external services. For example, if
    // we wanted to return the stores the product is available in from
    // a custom data store.
    var dbProduct = DbContext.Products.GetBySku(id);
    var shopModel = new Product(CurrentPage)
    {
        Sku = id,
        AvailableStores = dbProduct.AvailableStores
    };

    return View(shopModel);
}

This method is a bit more interesting, here we get some extra data from a different source, in this case a DbContext, but this can be anything you want, using the id we get from the route values. We use this extra data to create a custom model, wich includes the available stores, which we then render the view with.

It's important to note that this custom model must implement IPublishedContent, to do this we inherit from the ContentModel class, in this case our model looks like this:

public class Product : ContentModel
{
    public Product(IPublishedContent content) : base(content)
    {
    }

    public string Sku { get; set; }
    public IEnumerable<string> AvailableStores { get; set; }
}

What's great about this is that we can use this model as a type argument when inheriting from UmbracoViewPage in our model like so:

@inherits Umbraco.Cms.Web.Common.Views.UmbracoViewPage<RoutingDocs.Models.Product>

Which makes the model typed, so we can access the available stores like so:

<ul>
    @foreach (var store in Model.AvailableStores)
    {
        <li>@store</li>
    }
</ul>

But let's get back to our controller, the last thing we need now is to implement FindContent method so we can find content for our actions and serve it to them. First we need to be able to get our content, and properties, so we need to inject IUmbracoContextAccessor and IPublishedValueFallback and save them to some fields like so:

private readonly IUmbracoContextAccessor _umbracoContextAccessor;
private readonly IPublishedValueFallback _publishedValueFallback;

public ShopController(
    ILogger<UmbracoPageController> logger,
    ICompositeViewEngine compositeViewEngine,
    IUmbracoContextAccessor umbracoContextAccessor,
    IPublishedValueFallback publishedValueFallback)
    : base(logger, compositeViewEngine)
{
    _umbracoContextAccessor = umbracoContextAccessor;
    _publishedValueFallback = publishedValueFallback;
}

Now that we have our dependencies, and our action methods, we're finally ready to implement the FindContent method:

public IPublishedContent FindContent(ActionExecutingContext actionExecutingContext)
{
   if (_umbracoContextAccessor.TryGetUmbracoContext(out var umbracoContext))
   {
        var productRoot = umbracoContext.Content.GetById(2074);
        if (productRoot!=null)
        {

            if (actionExecutingContext.ActionDescriptor is ControllerActionDescriptor controllerActionDescriptor)
            {
                // Check which action is executing
                switch (controllerActionDescriptor.ActionName)
                {
                    case nameof(Index):
                        return productRoot;

                    case nameof(Product):
                        // Get the SKU/Id from the route values
                        if (actionExecutingContext.ActionArguments.TryGetValue("id", out var sku))
                        {
                            return productRoot
                                .Children
                                .FirstOrDefault(c => c.Value<string>(_publishedValueFallback, "sku") == sku.ToString());
                        }
                        else
                        {
                            return productRoot;
                        }
                 }
            }

            return productRoot;
        }
    }

    return null;
}

We start off by getting our product root using the UmbracoContext to get it based off its id. Next we need to figure out what action is being requested. To do this we cast the actionExecutingContext.ActionDescriptor to a ControllerActionDescriptor and use its ActionName propperty. If the action name is index, we return the product root. If it's product, we try to get the SKU from the route value id. Then we try to find the child node which matches the SKU and return that.

There's only one last thing to do. We need to register our shop controller. If you're creating a controller for your own site you can do it in the Program.cs like so:

app.UseUmbraco()
    .WithMiddleware(u =>
    {
        u.UseBackOffice();
        u.UseWebsite();
    })
    .WithEndpoints(u =>
    {
        u.EndpointRouteBuilder.MapControllerRoute(
                "Shop Controller",
                "/shop/{action}/{id?}",
                new {Controller = "Shop", Action = "Index"});

        u.UseBackOfficeEndpoints();
        u.UseWebsiteEndpoints();
    });

As you can see there's nothing Umbraco specific abouth the controller routing, it's using the default MapController route of the EndpointRouteBuilder, we give our mapping a name, a pattern for the controller and some default values, so if no action is specified it will default to Index.

If you're creating a package you won't have access to the Program.cs, so instead you can use a composer with an UmbracoPipelineFilter like so:

using Microsoft.AspNetCore.Builder;
using Microsoft.Extensions.DependencyInjection;
using Umbraco.Cms.Core.Composing;
using Umbraco.Cms.Core.DependencyInjection;
using Umbraco.Cms.Web.Common.ApplicationBuilder;

namespace RoutingDocs.Controllers;

public class ShopControllerComposer : IComposer
{
    public void Compose(IUmbracoBuilder builder)
    {
        builder.Services.Configure<UmbracoPipelineOptions>(options =>
        {
            options.AddFilter(new UmbracoPipelineFilter(nameof(ShopController))
            {
                Endpoints = app => app.UseEndpoints(endpoints =>
                {
                    endpoints.MapControllerRoute(
                        "Shop Controller",
                        "/shop/{action}/{id?}",
                        new {Controller = "Shop", Action = "Index"});
                })
            });
        });
    }
}

With that we have our controller with a custom route within an Umbraco context.

Client-Side Requests

If the endpoint of your custom route is considered a client-side request e.g. /sitemap.xml, you will need to make a few changes to get this to work.

Define your route as before, specifying the correct client type route:

.WithEndpoints(u =>
{
    u.EndpointRouteBuilder.MapControllerRoute("Sitemap Xml", "/sitemap.xml",
        new { Controller = "SitemapXml", Action = "Index" });
});

You will need to configure your route request options within your Program.cs class. For single routes:

builder.Services.Configure<UmbracoRequestOptions>(options =>
{
    options.HandleAsServerSideRequest = httpRequest => httpRequest.Path.StartsWithSegments("/sitemap.xml");
});

Or it can handle multiple routes:

builder.Services.Configure<UmbracoRequestOptions>(options =>
{
    string[] allowList = new[] {"/sitemap.xml", ...};
    options.HandleAsServerSideRequest = httpRequest =>
    {
        foreach (string route in allowList)
        {
            if (httpRequest.Path.StartsWithSegments(route))
            {
                return true;
            }
        }

        return false;
    };
});

In your FindContent method you should still be able to access and use IUmbracoContextAccessor through standard DI:

public IPublishedContent? FindContent(ActionExecutingContext actionExecutingContext)
{
    IUmbracoContext context = _umbracoContextAccessor.GetRequiredUmbracoContext();
    IPublishedContent? content = context.Content?.GetAtRoot().FirstOrDefault();

    return content;
}

Attribute routing with IVirtualPageController

One of the benefits of the IVirtualPageController is that it allows you to use attribute routing. If you wish to use attribute routing you must use an IVirtualPageController and decorate your controller and/or actions with the Route attribute. If we want to convert our above example into using attribute routing we must first add the attributes to our actions:

[Route("[controller]")]
[Route("[controller]/[action]")]
[HttpGet]
public IActionResult Index()

[Route("[controller]/[action]/{id?}")]
[HttpGet]
public IActionResult Product(string id)

Now all we need to do is change our routing to use EndpointRouteBuilder.MapControllers(); instead of adding a specific route.

app.UseUmbraco()
    .WithMiddleware(u =>
    {
        u.UseBackOffice();
        u.UseWebsite();
    })
    .WithEndpoints(u =>
    {
        u.EndpointRouteBuilder.MapControllers();

        u.UseBackOfficeEndpoints();
        u.UseWebsiteEndpoints();
    });

This will give us routing that's similar to what we have in the other example. It's worth noting that there's no defaults when using attribute routing, so to allow our index action to be accessed through both /shop and /shop/index, we add two attributes, specifying both routes individually.

Custom route with ForUmbracoPage

Making a custom route within the Umbraco context using ForUmbracoPage is similar to using IVirtualPageController. The main difference is that with ForUmbracoPage we no longer find the content from within the controller. Instead we assign the FindContent method when routing the controller. One important thing about ForUmbracoPage is that attribute routing is not available. To make our example from above work with ForUmbracoPage, we want to remove any attribute routing, and no longer implement IVirtualPageController. We'll also remove the FindContent method, our controller will then end up looking like this:

using Microsoft.AspNetCore.Mvc;
using Microsoft.AspNetCore.Mvc.ViewEngines;
using Microsoft.Extensions.Logging;
using RoutingDocs.Models;
using RoutingDocs.Persistence;
using Umbraco.Cms.Web.Common.Controllers;

namespace RoutingDocs.Controllers;

public class ShopController : UmbracoPageController
{
    public ShopController(
        ILogger<UmbracoPageController> logger,
        ICompositeViewEngine compositeViewEngine)
        : base(logger, compositeViewEngine)
    { }

    [HttpGet]
    public IActionResult Index()
    {
        // CurrentPage (IPublishedContent) will be the content returned
        // from the FindContent method.

        // return the view with the IPublishedContent
        return View(CurrentPage);
    }

    [HttpGet]
    public IActionResult Product(string id)
    {
        // CurrentPage (IPublishedContent) will be the content returned
        // from the FindContent method.

        // One example of using a custom route would be to include additional
        // model information based on external services. For example, if
        // we wanted to return the stores the product is available in from
        // a custom data store.
        var dbProduct = DbContext.Products.GetBySku(id);
        var shopModel = new Product(CurrentPage)
        {
            Sku = id,
            AvailableStores = dbProduct.AvailableStores
        };

        return View(shopModel);
    }
}

As you can see we still inherit from UmbracoPageController to get access to the helper method CurrentPage, but the rest is a normal controller.

The Umbraco magic will now instead happen where we route the controller, here we will pass a Func<ActionExecutingContext, IPublishedContent> delegate to the ForUmbracoPage method, this delegate is then responsible for finding the content, for instance using a composer with the same logic as in the IVirtualPageController it will look like this:

public class ShopControllerComposer : IComposer
{
    public void Compose(IUmbracoBuilder builder)
    {
        builder.Services.Configure<UmbracoPipelineOptions>(options =>
        {
            options.AddFilter(new UmbracoPipelineFilter(nameof(ShopController))
            {
                Endpoints = app => app.UseEndpoints(endpoints =>
                {
                    endpoints.MapControllerRoute(
                        "Shop Controller",
                        "/shop/{action}/{id?}",
                        new {Controller = "Shop", Action = "Index"})
                        .ForUmbracoPage(FindContent); // Here we register our FindContent method
                })
            });
        });
    }

    private IPublishedContent FindContent(ActionExecutingContext actionExecutingContext)
    {
        // Resolve services from the container
        var umbracoContextAccessor = actionExecutingContext.HttpContext.RequestServices
            .GetRequiredService<IUmbracoContextAccessor>();
        var publishedValueFallback = actionExecutingContext.HttpContext.RequestServices
            .GetRequiredService<IPublishedValueFallback>();

        var umbracoContext = umbracoContextAccessor.GetRequiredUmbracoContext();
        var productRoot = umbracoContext.Content.GetById(2074);

        if (actionExecutingContext.ActionDescriptor is ControllerActionDescriptor controllerActionDescriptor)
        {
            // Check which action is executing
            switch (controllerActionDescriptor.ActionName)
            {
                case nameof(ShopController.Index):
                    return productRoot;

                case nameof(ShopController.Product):
                    // Get the SKU/Id from the route values
                    if (actionExecutingContext.ActionArguments.TryGetValue("id", out var sku))
                    {
                        return productRoot
                            .Children
                            .FirstOrDefault(c => c.Value<string>(publishedValueFallback, "sku") == sku.ToString());
                    }
                    else
                    {
                        return productRoot;
                    }
            }
        }
        return productRoot;
    }
}

The Compose method of our composer is much the same as any other normal routing. The one difference is that we call ForUmbracoPage on the MapControllerRoute where we pass in our FindContent method. The FindContent method is almost the same as it was in the controller in the IVirtualPageController example, with one important difference. Since we can no longer inject our required service into the constructor, we instead request them using actionExecutingContext.HttpContext.RequestServices.GetRequiredService. You should not save the HttpContext or the IServiceProvider you get from the actionExecutingContext to a field or property on the class. The reason for this is that they will be specific to each request.

With this we have a custom routed controller within the Umbraco pipeline. If you navigate to /shop or /shop/product/<SKU> you will see the controllers actions being called with the content found in FindContent.zspo

Middleware is responsible for processing/customizing an incoming request before generating an outgoing response. Shortly defined the middleware is a step to check the requests before giving a response back.

Umbraco automatically configures all required middleware in the WithMiddleware() method in a specific order based on the Microsoft documentation.

You can use Umbraco pipeline filters in case you want to add your own middleware before, in-between or after the default Umbraco middleware. Filters are added by configuring the UmbracoPipelineOptions and require an instance of IUmbracoPipelineFilter that contains the following callbacks:

• PrePipeline - executed before any Umbraco-specific middleware is added, an example can be URL rewrites.

• PreRouting - executed after the static files middleware and before the routing middleware is added (using UseRouting()). It can also be used to change the incoming URL.

• PostRouting - executed after the routing middleware is added and can be used to configure Cross-origin resource sharing (CORS).

• PostPipeline - executed after all Umbraco-specific middleware is added.

• Endpoints - executed right before the Umbraco-specific endpoints are added using WithEndpoints().

The addition of the PostRouting callback is to allow correctly configuring the Cross-Origin Resource Sharing (CORS) middleware without having to use the WithCustomMiddleware() method.

• IUmbracoPipelineFilter is an interface in Umbraco that allows the creation of custom filters which then modifies the behavior of the request pipeline. It can be used to change different aspects of how Umbraco handles incoming requests, such as changing content or adding security checks.

• WithCustomMiddleware() is a method that can be used in Umbraco for adding custom middleware. This includes some specific customizable instructions that run in the request processing pipeline.

Configuring the Cross-Origin Resource Sharing (CORS) middleware

Create a composer with the following:

using Umbraco.Cms.Core.Composing;
using Umbraco.Cms.Web.Common.ApplicationBuilder;

public class CorsComposer : IComposer
{
    public const string AllowAnyOriginPolicyName = nameof(AllowAnyOriginPolicyName);

    public void Compose(IUmbracoBuilder builder)
        => builder.Services
        .AddCors(options => options.AddPolicy(AllowAnyOriginPolicyName, policy => policy.AllowAnyOrigin()))
        .Configure<UmbracoPipelineOptions>(options => options.AddFilter(new UmbracoPipelineFilter("Cors", postRouting: app => app.UseCors())))
        // For testing only
        .Configure<UmbracoPipelineOptions>(options => options.AddFilter(new UmbracoPipelineFilter("CorsTest", endpoints: app => app.UseEndpoints(endpoints =>
        {
            endpoints.MapGet("/echo", context => context.Response.WriteAsync("echo")).RequireCors(AllowAnyOriginPolicyName);
            endpoints.MapGet("/echo2", context => context.Response.WriteAsync("echo2"));
        }))));
}

You should be able to request /echo from any origin, but get an error when trying to fetch /echo2 from a foreign origin. This can be tested by pasting the following JavaScript code in browser console when not on the local origin/Umbraco website (such as umbraco.com):

fetch('https://localhost:44331/echo', {method: 'GET'}).then(result => result.text().then(text => console.log(text)));
fetch('https://localhost:44331/echo2', {method: 'GET'}).then(result => result.text().then(text => console.log(text)));

This should return the following:

An additional bonus is that this doesn't require any changes in your Program.cs file. It also allows packages to enable Cross-Origin Resource Sharing (CORS) and configure their own policies.

Users that currently use WithCustomMiddleware() will need to add calls to RunPreRouting() and RunPostRouting(). This is similar to Umbraco adding additional middleware in future versions.

The followed method is called on the "PublishedContentRequest.PrepareRequest()" method: FindPublishedContentAndTemplate(). We discuss shortly what this method is doing:

1. FindPublishedContent ()

2. Handles redirects

3. HandlePublishedContent()

4. FindTemplate()

5. FollowExternalRedirect()

6. HandleWildcardDomains()

HandlePublishedContent

• No content?

• Run the LastChanceFinder

• Is an IContentFinder, resolved by ContentLastChanceFinderResolver

• By default, is null (= ugly 404)

• Follow internal redirects

• Take care of infinite loops

• Ensure user has access to published content

• Else redirect to login or access denied published content

• Loop while there is no content

• Take care of infinite loops

FindTemplate

• Use altTemplate if

• Initial content

• Internal redirect content, and InternalRedirectPreservesTemplate is true

• No alternate template?

• Use the current template if one has already been selected

• Else use the template specified for the content, if any

• Alternate template?

  • Use the alternate template, if any

  • Else use what’s already there: a template, else none

• Alternate template is used only if displaying the intended content

  • Except for internal redirects

  • If you enable InternalRedirectPreservesTemplate

  • Which is false by default

• Alternate template replaces whatever template the finder might have set

  • ContentFinderByNiceUrlAndTemplate

  • /path/to/page/template1?altTemplate=template2  template2

• Alternate template does not falls back to the specified template for the content

  • /path/to/page?altTemplate=missing  no template

  • Even if the page has a template

• But preserves whatever template the finder might have set

  • /path/to/page/template1?altTemplate=missing  template1

FollowExternalRedirect()

• Content.GetPropertyValue("umbracoRedirect")

• If it’s there, sets the published content request to redirect to the content

• Will trigger an external (browser) redirect

HandleWildcardDomains()

• Finds the deepest wildcard domain between

• Domain root (or top)

• Request’s published content

• If found, updates the request’s culture accordingly

This implements separation between hostnames and cultures

Use a custom controller to handle and control incoming requests for content pages based on a specific Document Type

What is Umbraco Route Hijacking?

By default, all front-end requests to an Umbraco site are auto-routed via the 'Index' action of a core Controller: Umbraco.Cms.Web.Common.Controllers.RenderController. This core controller handles the incoming request, builds the associated PublishedContent model, and passes this to the appropriate Umbraco Template/MVC View.

It is however possible to implement a custom Controller to replace this default implementation to give complete control over this execution.

For example:

• To enrich the view model passed to the template with additional properties (from other published content items or outside Umbraco)

• To implement serverside paging

• To implement any custom/granular security

• To return alternative templates depending on some custom business logic

This replacement of the default controller can be made 'globally' for all requests (see last example). It can also be by 'hijacking' requests for types of pages based on their specific Document Type following this controller naming convention: [DocumentTypeAlias]Controller.

Creating a custom controller

Example: Hijacking route requests to a 'product' page

In the following example, imagine an Umbraco site with a set of 'product' pages created from a Document Type called 'Product Page' with an alias 'productPage'.

Create a custom locally declared controller in the Umbraco web application project named 'ProductPageController'.

Ensure that this controller inherits from the base controller Umbraco.Cms.Web.Common.Controllers.RenderController.

eg:

All requests to any product pages in the site will be hijacked and routed through the custom ProductPageController.

If you prefer to use an async controller your need to override both the sync and the async Index()-methods. This is done to disable the default behavior from the base controller.

This example shows the default behavior that Umbraco's core RenderController provides. The 'Index' action of the controller is executed, and the CurrentTemplate helper sends the model containing the details of the published content item related to the request to the relevant template/view.

Routing via template

A further convention is that if an action on the controller has a name that matches the template name, this action will be executed instead of the default 'Index' action.

Example: Hijacking route requests to a 'product' for an alternative 'AMP' template

In this example, the Product Page Document Type has two templates 'ProductPage' and 'ProductAmpPage'. We can hijack and handle the requests to the two templates differently.

Create the Controller as before:

How can a page be requested via two different templates?

The page in Umbraco will have a single 'template' selected as it's default template, but it's possible to call this same page on a different template by adding ?altTemplate=othertemplatename to the Url QueryString eg:

/products/superfancyproduct/?altTemplate=ProductAmpPage

Summary - How the route hijacking convention works

• Document Type name = controller name

• Template name = action name (if no action matches or is not specified - then the 'Index' action will be executed).

• Controller Inherits from Umbraco.Cms.Web.Common.Controllers.RenderController

Returning a view with a custom model

The steps to achieve this will differ, depending if your template views are using IPublishedContent or Modelsbuilder generated Models.

Changing the @inherits directive of your template

By default, your Umbraco Template will be based on the ContentModel that the default RenderController passes through to it.

The default inherits statement:

or if you are using modelsbuilder:

<> contains a model generated for each Document Type to give strongly typed access to the Document Type properties in the template view.

To use a specific custom view model, the @inherits directive will need to be updated to reference your custom model using the Umbraco.Cms.Web.Common.Views.UmbracoViewPage<T> format where 'T' is the type of your custom model.

So for example, if your custom model is of type 'MyProductViewModel' then your @inherits directive will look like:

In most cases you will need your custom model to build upon the underlying existing PublishedContent model for the page. This can be achieved by making your custom model inherit from a special base class called PublishedContentWrapped:

PublishedContentWrapped will take care of populating all the usual underlying Umbraco properties and means the @Model. syntax will continue to work in the layouts used by your template.

Using Modelsbuilder you will find that all the generated models have a constructor that takes an IPublishedContent item in a similar way:

We can now populate our custom view model in our controller and use the values from the custom model in our template view:

and in our template

Processing QueryString values in the controller

You can also pass values directly into the controller action using the query string.

The values in the querystring will be bound to the matching parameters defined in the controller's action:

Controller Injection

For example:

To wire up a concrete instance of IMadeUpProductService, use a composer:

Replace Umbraco's default RenderController

You can replace Umbraco's default implementation of RenderController with your own custom controller for all MVC requests. This is possible by assigning your own default controller type in the Umbraco setup during initialization.

You can achieve this by updating the options for UmbracoRenderingDefaultsOptions in Program.cs.

First of all, you have to create your own controller. Your custom implementation of RenderController should either inherit from the core RenderController as in the examples above or implement the IRenderController interface.

Implement the IRenderController:

Or inherit from RenderController

The last step is to configure Umbraco to use your implementation. You can do that in the Program.cs class.

User Overview

Whenever a document is published, and this causes changes to its URL (and any of its descendants' URLs), Umbraco makes a note of the old URLs. Whenever an incoming request is served and the default content finders cannot find a matching published document, Umbraco checks whether the URL matches one of these saved URLs. If a match is found, Umbraco returns a "301 Redirect" response pointing to the new URL of the document.

The URL Redirect Management functionality does not support rewriting "rules" (e.g. regular expressions), nor complex scenarios (e.g. changing the culture and hostnames configuration). There are already powerful solutions to deal with these types of situations, such as Microsoft's own module for IIS. Since netcore is decoupled from the webserver hosting it, your approach for URL rewriting, outside what Umbraco provide out of the box, will depend on what you use to host your solutions, but for more info on the IIS Url Rewrite module have a look at the .

Dashboard

It is possible to list the redirect URLs via the Redirect Url Management dashboard in the Content section. This dashboard lists the original URL, new URL, and culture. It also allows you to delete a URL redirect.

In addition, the dashboard can be used to disable or enable the 301 Redirect Management (via the appsettings.json configuration option described below - note that this requires an application restart to take effect).

Technical Overview

Anytime a document is published and its corresponding url segment changes, Umbraco checks its URL (and all its descendants' URLs) for changes. For every URL that has changed, it creates (or updates) a row in the umbracoRedirectUrl table. Rows in this table contain: the old url, the create date, and the target content identifier, culture, and a url hash.

Umbraco registers a new content finder, ContentFinderByRedirectUrl, which runs as a normal content finder after the other content finders. It looks for the incoming URL in the database table and, if found, computes the URL of the target document and returns a "301 Redirect". These redirects are considered "permanent". It's good to note that we explicitly set no-cache headers on these redirects so that when they change, browsers update the URL immediately. They are a "true" 301, however, and search engines will accept them as such.

Enable / Disable / Configure

The 301 Redirect Management feature is enabled by default.

It is possible to disable the feature entirely (both generating URLs in the database table, and running the content finder) by editing the appsettings.json file:

To create a custom content finder, with custom logic to find an Umbraco document based on a request, implement the IContentFinder interface:

and use either an Umbraco builder extension, or a composer to add it to it to the ContentFindersCollection.

Umbraco runs all content finders in the collection 'in order', until one of the IContentFinders returns true. Once this occurs, the request is then handled by that finder, and no further IContentFinders are executed. Therefore the order in which ContentFinders are added to the ContentFinderCollection is important.

The ContentFinder can set the PublishedContent item for the request, or template or even execute a redirect.

Example

This IContentFinders will find a document with id 1234, when the Url begins with /woot.

Adding and removing IContentFinders

You can use either an extension on the Umbraco builder or a composer to access the ContentFinderCollection and add or remove specific ContentFinders.

Umbraco builder extension

First create the extension method:

Then invoke in the Program.cs file:

Composer

NotFoundHandlers

To set your own 404 finder create an IContentLastChanceFinder and set it as the ContentLastChanceFinder. (perhaps you have a multilingual site and need to find the appropriate 404 page in the correct language).

A IContentLastChanceFinder will always return a 404 status code. This example creates a new implementation of the IContentLastChanceFinder and gets the 404 page for the current language of the request.

You can configure Umbraco to use your own implementation in the Program.cs file:

All about Umbraco's routing pipeline & the types of Controllers used in Umbraco, how they work, and what they are used for.

Explains how Umbraco builds its URLs and how the URLs are mapped back to content items.

Describes the Umbraco special/reserved Property Type aliases that can be used to directly manipulate Umbraco's default routing pipeline. These special Property Type aliases can be useful when creating an Umbraco website.

What is a Surface Controller and how to use them?

What is an API Controller and how to use them?

Creating custom controllers to have 100% full control over how your pages are rendered. Also known as: Hijacking Umbraco Routes

How to specify your own custom MVC routes in your Umbraco application?

Routing requirements for authenticated controllers for both front-end and the backoffice.

Moving and renaming Umbraco documents will lead to URL redirects to be created.

Umbraco Training

Umbraco HQ offers a full-day training course covering advanced routing, custom page controllers and models, and backoffice integration. The course targets backend ASP.NET MVC developers working with service and application integration on Umbraco.

This section describes what the Umbraco Request Pipeline is. It explains how Umbraco matches a document to a given request and how it generates a URL for a document.

Request pipeline

What is the pipeline

The request pipeline is the process of building up the URL for a node and resolving a request to a specified node. It ensures that the right content is sent back.

Outbound vs Inbound

Customizing the pipeline

Is started in UmbracoRouteValueTransformer where it gets the HttpContext and RouteValueDictionary from the netcore framework:

 async ValueTask<RouteValueDictionary> TransformAsync(…)

What it does:

• It ensures Umbraco is ready, and the request is a document request.

• Ensures there's content in the published cache, if there isn't it routes to the RenderNoContentController which displays the no content page you see when running a fresh install.

• Creates a published request builder.

• Routes the request with the request builder using the PublishedRouter.RouteRequestAsync(…).

  • This will handle redirects, find domain, template, published content and so on.

  • Build the final IPublishedRequest.

• Sets the routed request in the Umbraco context, so it will be available to the controller.

• Create the route values with the UmbracoRouteValuesFactory.

  • This is what actually routes your request to the correct controller and action, and allows you to hijack routes.

• Set the route values to the http context.

• Handles posted form data.

• Returns the route values to netcore so it routes your request correctly.

RouteRequestAsync

When the RouteRequestAsync method is invoked on the PublishedRouter it will:

• FindDomain().

• Handle redirects.

• Set culture.

• Find the published content.

  • Only if it doesn't exist, allowing you to handle it in a custom way with a custom router handler.

• Find the template.

• Set the culture (again, in case it was changed).

• Publish RoutingRequestNotification.

• Handle redirects and missing content.

• Initialize a few internal stuff.

We will discuss a few of these steps below.

FindDomain()

The FindDomain method looks for a domain matching the request Uri

• Using a greedy match: “domain.com/foo” takes over “domain.com”.

• Sets published content request’s domain.

• If a domain was found.

  • Sets published content request’s culture accordingly.

  • Computes domain Uri based upon the current request ("domain.com" for "http://domain.com" or "https://domain.com").

• Else.

  • Sets published content request’s culture by default (first language, else system).

Find published content

When finding published content the PublishedRouter will first check if the PublishedRequestBuilder already has content, if it doesn't the content finders will kick in. There a many different types of content finders, such as find by url, by id path, and more. If none of the content finders manages to find any content, the request will be set as 404, and the ContentLastChanceFinder will run, this will try to find a page to handle a 404, if it can't find one, the ugly 404 will be used.

You can also implement your own content finders and last chance finder, for more information, see IContentFinder

The PublishedRouter will also follow any internal redirects there might be, it is however limited, as to not spiral out of control if there is an infite loop of redirects.

Find template

Once the content has been found, the PublishedRouter moves on to finding the template.

First off it checks if any content was found, if it wasn't it sets the template to null, since there can't be a template without content.

Next it checks to see if there is an alternative template which should be used. An alternative template will be used if the router can find a value with the key "altTemplate", in either the querystring, form, or cookie, and there is content found by the contentfinders, so not the 404 page, or it's an internal redirect and the web routing setting has InternalRedirectPreservesTemplate.

If no alternative template is found the router will get the template with the file service, using the ID specified on the published content, and then assign the template to the request.

If an alternative template is specified, the router will check if it's an allowed template for the content, if the template is not allowed on that specific piece of content it will revert to using the default template. If the template is allowed it will then use the file service to get the specified alternative template and assign the template to the request.

Redirects

The router will pick up the redirect and redirect. There is no need to write your own redirects:

public class PublishedRequestHandler : INotificationHandler<RoutingRequestNotification>
{
    public void Handle(RoutingRequestNotification notification)
    {
        var requestBuilder = notification.RequestBuilder;
        var content = requestBuilder.PublishedContent;
        var redirect = content.Value<string>("myRedirect");
        if (!string.IsNullOrWhiteSpace(redirect))
        {
            requestBuilder.SetRedirect(redirect);
        }
    }
}

Missing template?

In case the router can't find a template, it will try and verify if there's route hijacking in place, if there is, it will run the hijacked route. If route hijacking is not in place, the router will set the content to null, and run through the routing of the request again, in order for the last chance finder to find a 404.

The inbound process is triggered by UmbracoRouteValueTransformer and then handled with the Published router. The published content request preparation process kicks in and creates a PublishedRequestBuilder which will be used to create a PublishedContentRequest.

The PublishedContentRequest object represents the request which Umbraco must handle. It contains everything that will be needed to render it. All this occurs when the Umbraco modules knows that an incoming request maps to a document that can be rendered.

public class PublishedContentRequest
{
  public Uri Uri { get; }
  …
}

There are 3 important properties, which contains all the information to find a node:

public bool HasDomain { get; }
public DomainAndUri Domain { get; }
public CultureInfo Culture { get; }

Domain is a DomainAndUri object that is a standard Domain plus the fully qualified uri. For example, the Domain may contain "example.com" whereas the Uri will be fully qualified for example "https://example.com/".

It contains the content to render:

public bool HasPublishedContent { get; }
public IPublishedContent PublishedContent { get; set; }
public bool IsInternalRedirect { get; }
public bool IsRedirect {get; }

Contains template information:

public bool HasTemplate { get; }
public string GetTemplateAlias { get; }
public ITemplate Template {get; }

The published request is created using the PublishedRequestBuilder, which implements IPublishedRequestBuilder. It's only in this builder that it's possible to set values, such as domain, culture, published content, redirects, and so on.

You can subscribe to the 'routing request' notification, which is published right after the PublishedRequestBuilder has been prepared, but before the request is built, and processed. Here you can modify anything in the request before it is built and processed! For example content, template, etc:

using Umbraco.Cms.Core.Events;
using Umbraco.Cms.Core.Notifications;

namespace Umbraco9.NotificationHandlers;

public class PublishedRequestHandler : INotificationHandler<RoutingRequestNotification>
{
    public void Handle(RoutingRequestNotification notification)
    {
        var requestBuilder = notification.RequestBuilder;
        // Do something with the IPublishedRequestBuilder here 
    }
}

A surface controller is an MVC controller that interacts with the front-end rendering of an Umbraco page. They can be used for rendering view components and for handling form data submissions. Surface controllers are auto-routed, meaning that you don't have to add/create your own routes for these controllers to work.

What is a surface controller?

It is a regular ASP.NET Core MVC controller that:

• Is auto-routed, meaning you don't have to setup any custom routes to make it work

• Is used for interacting with the front-end of Umbraco (not the backoffice)

Since any surface controller inherits from the Umbraco.Cms.Web.Website.Controllers.SurfaceController class, the controller instantly supports many of the helper methods and properties that are available on the base SurfaceController class including UmbracoContext. Therefore, all surface controllers have native Umbraco support for:

• Interacting with Umbraco routes during HTTP POSTs (i.e. return CurrentUmbracoPage(); )

• Rendering forms in Umbraco (i.e. @using (Html.BeginUmbracoForm<MyController>(...)){} )

• Rendering of ASP.NET Core MVC view components

Creating a surface controller

Surface controllers are plugins, meaning they are found when the Umbraco application boots. There are 2 types of surface controllers: locally declared & plugin based. The main difference between the two is that a plugin based controller gets routed via an MVC area, which is defined in the controller (see below). Because a plugin based controller is routed via an MVC area, it means that the views can be stored in a custom folder specific to the package it is being shipped in. This can be done without interfering with the local developer's MVC view files.

Locally declared controllers

A locally declared surface controller is one that is not shipped within an Umbraco package. It is created by the developer of the website they are creating. If you are planning on shipping a surface controller in an Umbraco package, then you will need to create a plugin based surface controller (see the next heading).

To create a locally declared surface controller:

• Create a controller that inherits from Umbraco.Cms.Web.Website.Controllers.SurfaceController

• The controller must be a public class.

• The controller must call the base constructor of SurfaceController

• The controller's name must be suffixed with the term Controller

• The controller must be inside a namespace

For example:

using Microsoft.AspNetCore.Mvc;
using Umbraco.Cms.Core.Cache;
using Umbraco.Cms.Core.Logging;
using Umbraco.Cms.Core.Routing;
using Umbraco.Cms.Core.Services;
using Umbraco.Cms.Core.Web;
using Umbraco.Cms.Infrastructure.Persistence;
using Umbraco.Cms.Web.Website.Controllers;

namespace RoutingDocs.Controllers;

public class MyController : SurfaceController
{
    public MyController(
        IUmbracoContextAccessor umbracoContextAccessor,
        IUmbracoDatabaseFactory databaseFactory,
        ServiceContext services,
        AppCaches appCaches,
        IProfilingLogger profilingLogger,
        IPublishedUrlProvider publishedUrlProvider)
        : base(umbracoContextAccessor, databaseFactory, services, appCaches, profilingLogger, publishedUrlProvider)
    {
    }

    public IActionResult Index()
    {
        return Content("Hello world");
    }
}

Routing for locally declared controllers

All locally declared controllers gets routed to:

/umbraco/surface/{controllername}/{action}/{id}

They do not get routed via an MVC area, so any views must exist in the following folders:

• /Views/{controllername}/

• /Views/Shared/

• /Views/

Plugin based controllers

If you are shipping a surface controller in a package, then you should definitely be creating a plugin based surface controller. The only difference between creating a plugin based controller and locally declared controller, is that you need to add an attribute to your class, which defines the MVC area you'd like your controller to be routed through. Here's an example:

using Microsoft.AspNetCore.Mvc;
using Umbraco.Cms.Core.Cache;
using Umbraco.Cms.Core.Logging;
using Umbraco.Cms.Core.Routing;
using Umbraco.Cms.Core.Services;
using Umbraco.Cms.Core.Web;
using Umbraco.Cms.Infrastructure.Persistence;
using Umbraco.Cms.Web.Common.Attributes;
using Umbraco.Cms.Web.Website.Controllers;

namespace SurfaceControllerPackage;

[PluginController("SurfaceControllerPackage")]
public class MyController : SurfaceController
{
    public MyController(
        IUmbracoContextAccessor umbracoContextAccessor,
        IUmbracoDatabaseFactory databaseFactory,
        ServiceContext services,
        AppCaches appCaches,
        IProfilingLogger profilingLogger,
        IPublishedUrlProvider publishedUrlProvider)
        : base(umbracoContextAccessor, databaseFactory, services, appCaches, profilingLogger, publishedUrlProvider)
    {
    }

    public IActionResult Index()
    {
        return Content("Hello world");
    }
}

In the above, the surface controller will belong to the MVC area called 'SurfaceControllerPackage'. Perhaps it is obvious, but if you are creating a package that contains many surface controllers, then you should most definitely ensure that all of your controllers are routed through the same MVC area.

Routing for plugin based controllers

All plugin based controllers get routed to:

/umbraco/{areaname}/{controllername}/{action}/{id}

Since they get routed via an MVC area, your views should be placed in the following folder:

• ~/App_Plugins/{areaname}/Views/{controllername}/

• ~/App_Plugins/{areaname}/Views/Shared/

Since you're only able to place static filese within your package's App_Plugin folder, it's highly recommend to ensure that the area you use is the same as your package name, since that allows your views to be found.

The controller itself should not be placed in the App_Plugins folder, the App_Plugins folder is for static files only, compiled files like the controller will be included in the dlls used by the nuget package.

Protecting surface controller routes

If you only want a surface controller action to be available when it's used within an Umbraco form and not from the auto-routed URL, you can add the [ValidateUmbracoFormRouteString] attribute to the action method. This can be especially useful for plugin based controllers, as this makes sure the actions can only be activated from a form whenever it's used within the website.

namespace RoutingDocs.Controllers;

public class MyController : SurfaceController
{
    public MyController(
        IUmbracoContextAccessor umbracoContextAccessor,
        IUmbracoDatabaseFactory databaseFactory,
        ServiceContext services,
        AppCaches appCaches,
        IProfilingLogger profilingLogger,
        IPublishedUrlProvider publishedUrlProvider)
        : base(umbracoContextAccessor, databaseFactory, services, appCaches, profilingLogger, publishedUrlProvider)
    {
    }

    [HttpPost]
    [ValidateUmbracoFormRouteString]
    public IActionResult HandleSubmit()
    {
        return RedirectToCurrentUmbracoPage();
    }
}

Whenever you render an Umbraco form within your view using Html.BeginUmbracoForm<MyController>(...), the forms action will be the URL of the current page (not the auto-routed URL of the surface controller). Umbraco will therefore add a hidden ufprt field to the form with an encrypted value containing the controller, action and optional area (known as the 'Umbraco form route string'). On form submission, this value is decrypted and Umbraco will activate the specified action of the surface controller.

@using (Html.BeginUmbracoForm<MyController>("HandleSubmit"))
{
    <input type="submit" />
}

Preventing Cross-Site Request Forgery (XSRF/CSRF) Attacks

Cross-Site Request Forgery (CSRF) is an attack that forces an end user to execute unwanted actions on a web application in which they are currently authenticated.

If the token is not added automatically, for instance, if you don't use Html.BeginUmbracoForm or use an overload to Html.BeginForm where you've set the antiForgery parameter to false, you can add it manually like so:

@using (Html.BeginForm(nameof(ContactFormController.Submit), "ContactForm", new object(), FormMethod.Post, false /* this is where you disable it */, new object()))
{
  @Html.AntiForgeryToken()
  // Add your form fields here
}

If you are using a SurfaceController the antiforgery token will automatically be validated. However, if you are using a standard (non-umbraco) controller, you can manually specify it with the ValidateAntiForgeryToken attribute:

[HttpPost]
[ValidateAntiForgeryToken]
public ActionResult EditAction(FormViewModel formData)
{
    // Handle your form submit here
}

The BeginUmbracoForm and BeginForm will only add the antiforgery token to the form as a hidden input. This means that you have to manually handle this if you're sending the request via JavaScript, for example, ajax.

The routing expects the antiforgery token to be in a header called RequestVerificationToken. You can use the beforeSend hook to read the antiforgery token and set it as a header if you're using ajax:

$.ajax({
    beforeSend: function (xhr) {
        xhr.setRequestHeader("RequestVerificationToken",
        $('input:hidden[name="__RequestVerificationToken"]').val());
    }
});

For more information, see the Antiforgery in ASP.NET Core article.

Asynchronous surface controller actions

Surface controller actions can be asynchronous. A common naming convention for asynchronous methods is using an Async suffix for the action name. However, this will not work by default due to the inner workings of ASP.NET Core MVC.

Consider the following asynchronous surface controller action:

public class MySurfaceController : SurfaceController
{
    // ...

    [HttpPost]
    [ValidateAntiForgeryToken]
    public async Task<IActionResult> SomeMethodAsync()
    {
        // do some async work here
        return CurrentUmbracoPage();
    }
}

To use this action in a view you can add this:

@using (Html.BeginUmbracoForm<MySurfaceController>(nameof(MySurfaceController.SomeMethodAsync), FormMethod.Post))
{
 <button type="submit">Do the async work</button>
}

But once you click the button, you will encounter an error message along the lines of: InvalidOperationException: Could not find a Surface controller route in the RouteTable for controller name MySurface.

To counter this you need to instruct ASP.NET Core MVC to explicitly accept the Async suffix for controller names in the program.cs:

builder.Services.AddMvc(options =>
{
    options.SuppressAsyncSuffixInActionNames = false; 
});

Alternatively you can rename your surface controller action so it does not contain the Async suffix.

Surface Controller Actions

You can read more about the surface controller action result helpers.

This article describes how to work with API controllers in Umbraco projects. It focuses on creating REST services using ASP.NET Core-based API controllers.

API Overview

To better understand the basics of APIs, you can see the Microsoft ASP.NET Core API documentation. The documentation provides a solid foundation for API concepts in .NET environments..

Public APIs in Umbraco

Public APIs in Umbraco are similar to standard ASP.NET Core APIs. Below is an example of how to create an API in Umbraco:

using Microsoft.AspNetCore.Mvc;

namespace UmbracoDocs.Samples;

[ApiController]
[Route("/api/shop/products")]
public class ProductsController : Controller
{
    [HttpGet]
    public IActionResult GetAll() => Ok(new[] { "Table", "Chair", "Desk", "Computer" });
}

Adding Member Protection to Public APIs

You can secure your public APIs using front-end membership protection with the [UmbracoMemberAuthorize] attribute. This attribute allows you to restrict access based on member types, groups, or specific member IDs.

The available parameters are:

• AllowType: A comma-delimited list of allowed member types.

• AllowGroup: A comma-delimited list of allowed member groups.

• AllowMembers: A comma-delimited list of allowed member IDs.

To allow all members, apply the [UmbracoMemberAuthorize] attribute without parameters.

You can apply these attributes either at the controller level or at the action level.

Examples of Member Protection

The [UmbracoMemberAuthorize] attribute offers flexible options for securing your public APIs in Umbraco. The following examples show different ways to apply member protection, such as how to restrict access by member type, group, or specific IDs.

Example 1: Allow All Logged-In Members

In this example, any logged in member can access all actions in the ProductsController controller:

using Microsoft.AspNetCore.Mvc;
using Umbraco.Cms.Web.Common.Filters;

namespace UmbracoDocs.Samples;

[ApiController]
[Route("/api/shop/products")]
[UmbracoMemberAuthorize]
public class ProductsController : Controller
{
    [HttpGet]
    public IActionResult GetAll() => Ok(new[] { "Table", "Chair", "Desk", "Computer" });
}

Example 2: Restrict Access by Member Type

This example allows only logged-in members of type "Retailers" to access the GetAll action:

using Microsoft.AspNetCore.Mvc;
using Umbraco.Cms.Web.Common.Filters;

namespace UmbracoDocs.Samples;

[ApiController]
[Route("/api/shop/products")]
public class ProductsController : Controller
{
    [HttpGet]
    [UmbracoMemberAuthorize("Retailers", "", "")]
    public IActionResult GetAll() => Ok(new[] { "Table", "Chair", "Desk", "Computer" });
}