Twitter/X Space Downloader

readme.md

Twitter/X Space Downloader

Available here: https://gist.github.com/rodrigopolo/40029718a85963399784ae35b06adcaf

Dependencies:

• yt-dpl
• wget
• aria2c
• ffmpeg

Set the file permissions with chmod +x downloader.sh and run the script, with the URL of the Twitter/X Space as the first argument:

./downloader.sh https://twitter.com/i/spaces/1djxXNyNogvGZ

downloader.sh

#!/usr/bin/env bash

#
# Twitter/X Space Downloader Bash Script
# Copyright (c) 2024 Rodrigo Polo - rodrigopolo.com - The MIT License (MIT)
#

# Check if a stream URL is provided
if [ -z "$1" ]; then
	echo "Usage: $0 <stream_url>"
	exit 1
fi

# Modifying the internal field separator
IFS=$'\t\n'

SPACEURL="$1"

STREAM=$(yt-dlp --cookies-from-browser opera -g $SPACEURL)
FILE_NAME=$(yt-dlp --cookies-from-browser opera --get-filename -o "%(upload_date)s - %(uploader_id)s.%(title)s.%(id)s.%(ext)s" $SPACEURL)




# Get the stream path
STREAMPATH=$(echo "$STREAM" | grep -Eo "(^.*[\/])")

# Download the stream
if ! wget "$STREAM" -O stream.m3u8; then
	echo "Failed to download the stream."
	exit 1
fi

# Prefix the URLs for the chunks
cat stream.m3u8 | sed -E "s|(^[^.#]+.aac$)|$STREAMPATH\1|g" > modified.m3u8

# Download the chunks
aria2c -x 10 --console-log-level warn -i modified.m3u8

# Join the chunks
ffmpeg -i stream.m3u8 -vn -acodec copy -movflags +faststart "$FILE_NAME"

# Clean-up temporary files
eval "$(cat stream.m3u8| grep -Eo "(^[^.#]+.aac$)" | sed 's/^/rm /')"
rm stream.m3u8 modified.m3u8

BTW, you can also use yt-dlp, already supports it
https://github.com/yt-dlp/yt-dlp

It is a fork of youtube-dl, here is a cheat sheet
https://github.com/rodrigopolo/cheatsheets/blob/master/youtube-dl.md

Thanks! I’m looking for a way to transcribe the Twitter spaces next and then summarize the Twitter spaces using a local LLM like Ollama. Do you have any suggestions?

Thanks! I’m looking for a way to transcribe the Twitter spaces next and then summarize the Twitter spaces using a local LLM like Ollama. Do you have any suggestions?

1. Whisper: https://github.com/rodrigopolo/cheatsheets/blob/master/whisper.md
2. Adobe Photoshop
3. DaVinci Resolve
4. Subtitle Edit (PC Only) that uses Whisper or Vosk/Kaldi https://www.nikse.dk/subtitleedit

Thank you for the suggestions! Appreciate you.

The cheatsheets are really helpful.