Funcaster

Serverless .NET solution for hosting your podcasts with (nearly) zero costs using Azure Functions and Azure Storage Account.

Supported platforms

How does it work

To start publishing your podcast you simply need two things: Public accessible storage for your audio files, and RSS feed providing information necessary information about podcasts and episodes in valid XML format. Only if we would have cloud-based services providing terabytes of storage for a ridiculously low price and service for providing an Http-based serverless API with millions of monthly invocations for free... Oh, wait! We have Azure Storage Account & Azure Functions! Yes, this solution does exactly that - checking your blobs (audio files) on your Storage Account and providing valid RSS XML over Azure Functions.

Requirements

To use this solution you need to have access to Azure Portal and be logged in. If you don't have an account, you can create one for free.

Pre-installation

You need to have binaries ready for deployment. You can easily download the latest ZIP version from Releases, or you can build your own by checkout this repository and running:

> dotnet tool restore
> dotnet run publish

Custom build output, as well as ZIP file, can be found in the publish folder on the root level of the solution. No matter which one you'll choose, get your ZIP file ready for deployment described in the next chapter.

Installation

We will need to install and set up two things: Storage Account for your podcasts, and Azure Functions app for management around those files. It's pretty easy so just follow the steps and you are quickly done.

Storage Account

Create new Storage Account service

Once logged in Azure Portal, create a new Storage Account a wait for deployment is complete.

Create root container

All your podcast data need to be stored somewhere. This is why we create a new blob container called podcast with Blob permissions (anonymous read access for blobs only). This means that data in this folder can be accessed publicly (we need that).

Copy access key

To connect serverless part of Funcaster with newly created Storage Account, you need to copy a connection string.

Cool! Now everything is ready for the second part - Azure Functions.

Azure Functions

Create new Function App

Create a new Function app with any name you would like to use, select the same region your Storage Account is in, choose .NET6 and proceed to creation. If you would like to have built-in monitoring, turn on the Application Insights.

Setup connection string, runtime, and version

At this date (17. 10. 2021), Funcaster is built on .NET5 using dotnet-isolated as worker runtime, which is equivalent to ~3 extension version. We need to set this properly in configuration + add the connection string to the prepared Storage Account.

Deploy Funcaster

Ok, let's deploy our ZIP file! There are many ways (CI solutions like GitHub Actions or Azure DevOps pipelines are preferred), but let's take the shortcut here and use the KUDU ZIP deployment running at https://<NAME_OF_YOUR_FUNCTION_APP>.scm.azurewebsites.net/

Great! Your solution is almost ready! Now it's time to add some podcast content, can we?

Using Funcaster

Your podcasting solution is ready for serving files stored on blob storage. Let's start using it!!

Uploading new episodes

To add new episodes, you just simply upload files into the episodes (virtual) container in your podcast container:

Publishing episodes

After few seconds, Funcaster will find the newly uploaded file and create a basic metadata yaml file named the same as the original file. You can switch to a different view to see what's inside.

You can see that Funcaster sniffed the newly uploaded file and already prefilled basic information like url, type, or length for you. You need to give it some proper title (we don't wanna call our episodes as the empty string), description, and the duration of the episode in HH:MM:SS format. After updating those, just click on the Save button.

Great! We are nearly done! The last step is to set up our podcast metadata. That's why we have automatically created podcast.yaml file on the root level. Let's fill it, can we?

IT'S DONE! Let's see how our RSS feed looks like. Navigate to https://<NAME_OF_YOUR_FUNCTION_APP>.azurewebsites.net/rss and behold!

Using RSS feed

Now it's up to you where you wanna publish your podcast. I've already successfully published to Spotify Podcasts, Apple Podcasts and Google Podcasts but I'm pretty sure it will work for other providers too.

Should I use this instead of Soundcloud?

It's up to you. Soundcloud offers great social features like likes, a built-in web player, and many more, but it costs over 8 EUR / months. If you just need to publish your podcast for (almost) zero costs, Funcaster is here at your service.

What if I need an embedded player on my site?

Spotify and Apple offer an easy way of creating embedded players directly from the application or marketing tools. Or just use <audio> HTML5 tag.

What's the _index.yaml file for?

This file is automatically created and updated by Funcaster to be aware of all the episodes you have without going through the whole blob storage file system tree and scanning. You shouldn't care about this file until you need to...

What if I need to delete an episode?

Thanks to fact that BlobTrigger is not fired when something is removed, you need to do some manual work. First step is to remove YourFileToBeDeleted.mp3 and YourFileToBeDeleted.yaml. The second step is to go into _index.yaml and delete it from the index.

Can I have episodes stored in a deeper structure?

Sure! The only required thing is to have them stored in episodes folder - this is where Funcaster is sniffing around. But if you upload your file directly in this folder or you create episodes\some\completely\weird\logic\for\storing\my\files it's your call.