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

Works well! You may have to change the browser name in lines 19 and 20.

Works really well. Appreciate you sharing this to the public.

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.