Update README.md

README.md

@@ -44,7 +44,7 @@ The easiest way to install and use Hey Buddy is with [Anaconda/Miniconda](https:
 wget https://raw.githubusercontent.com/painebenjamin/hey-buddy/refs/heads/main/environment.yml
 conda env create -f environment.yml
 conda activate heybuddy
-pip install heybuddy
+pip install heybuddy[gpu]
 ```
 
 If you already have a working CUDA environment, you can just skip to `pip install heybuddy`. You will need to install `piper-phonemize` separately in this case, see [the rhasspy/piper-phonemizer releases page on GitHub](https://github.com/rhasspy/piper-phonemize/releases/) to find the latest wheel for your python version - for example, the environment file uses python 3.10, so it installs `https://github.com/rhasspy/piper-phonemize/releases/download/v1.1.0/piper_phonemize-1.1.0-cp310-cp310-manylinux_2_28_x86_64.whl`.
@@ -342,7 +342,7 @@ A number of evaluation metrics are recorded during training. The best way to acc
 **Hey Buddy!** is distributed as **two** JavaScript files that should be made available through any implementing web application:
 
 1. `hey-buddy.min.js` (90 kB) which provides the HeyBuddy API, and
-2. `hey-buddy-worklet.js` (1 kB) which implements the Audio Worklet interface which will be imported by the browser automatically.
+2. `hey-buddy-worklet.js` (1 kB) which implements the Audio Worklet interface which will be imported by the browser using the [AudioWorklet API](https://developer.mozilla.org/en-US/docs/Web/API/AudioWorklet).
 
 You need to make the ONNX runtime available prior to importing the HeyBuddy JS API. The easiest way to do this is to use a JS CDN, like so: