HLS & Nostr
My initial testing of using blossom as a backend for storing video segments & playlists has been successful, testing single resolution streams packaged as fMP4 segments.
The next step will be variable bandwidth streams so video can be flexible depending on the users connection speed. FFmpeg makes this easy by creating the video segments & playlists, only requiring me to map the Blossom hash to the playlist entries after they are uploaded.
Future plans:
- a native Android app, using ⚡ Dee Kay ⚡🇸🇪🇬🇧🇨🇿🇧🇷🇦🇹 (nprofile…m3sx) LCe library for compression & segmenting. Likely best for shorter videos due to hardware constraints.
- a command line app depending on ffmpeg & nak, useful for larger uploads with access to more powerful hardware & GPU acceleration.
I would like input from blossom server providers. With the segmenting aspect of this it will sidestep file size upload limits. I wonder though if the possible bandwidth savings would make this worth it, vs users uploading huge MP4 files and serving them with no optimization?
Ryan
npub1m6…9uaks
2026-04-12 16:50:55 UTC
Author Public Key
npub1m64hnkh6rs47fd9x6wk2zdtmdj4qkazt734d22d94ery9zzhne5qw9uaksPublished at
2026-04-12 16:50:55 UTCEvent JSON
{
"id": "58c36acc34462da6ddf2f28f0377284022de27fb8548cf5971b191dcaee3e9ff",
"pubkey": "deab79dafa1c2be4b4a6d3aca1357b6caa0b744bf46ad529a5ae464288579e68",
"created_at": 1776012655,
"kind": 1,
"tags": [
[
"alt",
"A short note: HLS \u0026 Nostr\n\nMy initial testing of using blossom a..."
],
[
"p",
"ca89cb11f1c75d5b6622268ff43d2288ea8b2cb5b9aa996ff9ff704fc904b78b",
"wss://nostr.mom/"
],
[
"client",
"Amethyst"
]
],
"content": "HLS \u0026 Nostr\n\nMy initial testing of using blossom as a backend for storing video segments \u0026 playlists has been successful, testing single resolution streams packaged as fMP4 segments. \n\nThe next step will be variable bandwidth streams so video can be flexible depending on the users connection speed. FFmpeg makes this easy by creating the video segments \u0026 playlists, only requiring me to map the Blossom hash to the playlist entries after they are uploaded.\n\nFuture plans:\n\n- a native Android app, using nostr:nprofile1qqsv4zwtz8cuwh2mvc3zdrl5853g365t9j6mn25edlul7uz0eyzt0zcpzpmhxue69uhkummnw3ezumt0d5hszxnhwden5te0wf5kymewv46jumn0wd68y6tp9eshqup0qyd8wumn8ghj7um9dejxjapwdehhxenvv9ex2tnrdakj7nrm3sx LCe library for compression \u0026 segmenting. Likely best for shorter videos due to hardware constraints.\n\n- a command line app depending on ffmpeg \u0026 nak, useful for larger uploads with access to more powerful hardware \u0026 GPU acceleration.\n\nI would like input from blossom server providers. With the segmenting aspect of this it will sidestep file size upload limits. I wonder though if the possible bandwidth savings would make this worth it, vs users uploading huge MP4 files and serving them with no optimization?\n\n",
"sig": "5f7fe1882015324ee5492b252fd04cf2202bd9184c2c394784891d7b64317a5b95822c6aa53357ec9beade8116b6d5bce31096bcc63574194b1059a8e7fe5b07"
}