Create speech-enabled apps with Azure Speech in Microsoft Foundry Tools: Review

Reaching the 'Exercise - Create a speech-enabled app' unit was a letdown; it's just a link to a GitHub Gist. You don't build the app. You don't run it. You just read the code for an app someone else built, then the quiz asks you what that code does. It's a reading comprehension test with extra steps.

I was hoping for a sandbox, like some of the other Azure AI modules, where you can play with the API for free in a controlled environment. No luck. To actually try anything shown here, you have to spin up a real Azure Speech resource, which means putting in a credit card. As a student, that's a non-starter for a one-hour module. I'm not risking a surprise bill just to try a text-to-speech endpoint.

So what you get for free is a high-level tour. You learn that you need a `SpeechConfig` object and can use `SpeechSynthesizer` for text-to-speech. You see some examples of SSML for adding pauses or changing voices, which is mildly interesting. The code snippets are clean and you could probably copy-paste them into your own project if you were already paying for Azure.

But as a learning experience? It's hollow. It doesn't explain the 'why' behind any of the async patterns or audio stream configurations. It just shows you the finished product. I got more practical experience from reading the quickstart docs for AssemblyAI and ElevenLabs, both of which have a free tier you can actually use without a credit card.