screenshot VS captureSystemAudio

No. It can't. I can capture anything on my screen with navigator.mediaDevides.getDisplayMedia({video: true}) alone https://github.com/guest271314/screenshot, or other means. If I wanted to, I can capture audio as well, with several different approaches, respectively https://github.com/guest271314/captureSystemAudio; and both in a single video, or multiple vides merged into one https://github.com/guest271314/MediaFragmentRecorder.

Synchronous. Though non-blocking when we read the message from the browser, execute the command, then stream data output from the local application to the browser piped from QuickJS std.popen(), capture_system_audio.js

I've already achieved the requirement multiple ways, already; from using Native Messaging https://github.com/guest271314/native-messaging-espeak-ng, to using GNU Core Utilities tail, to Deno.watchFs() https://github.com/guest271314/fs, et al., see captureSystemAudio. The one approach I have not yet achieved is compiling to Emscripten - with SSML support.

Sure, see https://github.com/guest271314/captureSystemAudio/blob/master/native_messaging/capture_system_audio/background.js, in pertinent part

This is my second time writing C source code capture_system_audio.c. I got a lot of help doing that, too . I basically implemented the same Native Messaging host algorithm in C, C++, Python, WASI/WebAssembly, and JavaScript (Node.js, QuickJS) which is the programming language I write code most using.

I just fetch the nightly node executable, get rid of everything else in the archive, use .mjs extension for import to work without package.json, then I can connect to the host https://github.com/guest271314/native-messaging-nodejs, and start subprocesses, e.g., raw PCM streams, where stdout is streamed to the client (browser), e.g, https://github.com/guest271314/captureSystemAudio/blob/master/native_messaging/capture_system_audio/capture_system_audio_node.js.

Technically I can just serve the raw data as long as the server is capable of opening creating sub-processes, e.g., using EventSource https://plnkr.co/edit/qX5tepJ38BuDOSZj?preview, in which case I would just need to convert this https://github.com/guest271314/captureSystemAudio/blob/master/native_messaging/capture_system_audio/capture_system_audio.c to accept GET request, read the URL, and send the named event stream to the client.

Chromium and Chrome refuses to capture monitor devices on Linux. Read this https://github.com/edisionnano/Screenshare-with-audio-on-Discord-with-Linux README and the README at https://github.com/guest271314/captureSystemAudio very carefully. At best you can capture tab audio, not whatever is output to speakears and headphones, in spite of the systemAudio constraints, which is poorly named, and actually misleading.

How is capturing speechSynthesis.speak() a security vector https://github.com/guest271314/captureSystemAudio?

Capture audio output to speakers or headphones to MP3 or Opus in WebM https://github.com/guest271314/captureSystemAudio/tree/master/native_messaging/capture_system_audio

An example of using Native Messaging to capture "What-U-Hear", sound output to speakers and headphone, launching native GUI's https://github.com/guest271314/captureSystemAudio/tree/master/native_messaging/capture_system_audio; an adapter for espeak-ng so we can capture the output of speech synthesis engine and stream output to the browser https://github.com/guest271314/native-messaging-espeak-ng.