i think i saw someone implement hls on blossom. if i did it again, that’s what id do. but that approach would still need some backend ffmpeg component to encode the audio 🤔