Convert a JSON response from Youtube transcript to Markdown

youtube_transcript_to_md.py

"""
Parses JSON response from https://www.youtube.com/youtubei/v1/get_transcript
and converts to Markdown document
"""

import json
import sys

def get_segments(data):
    return data['actions'][0]['updateEngagementPanelAction']['content']['transcriptRenderer']['content']['transcriptSearchPanelRenderer']['body']['transcriptSegmentListRenderer']['initialSegments']

def extract_transcript_segments(segments):
    if len(segments) == 0: return

    # Loop through subsequent content items and yield them
    for segment in segments:
        if 'transcriptSectionHeaderRenderer' in segment:
            text = segment['transcriptSectionHeaderRenderer']['sectionHeader']['sectionHeaderViewModel']['headline']['content']
            yield f"\n\n## {text}\n\n"

        if 'transcriptSegmentRenderer' in segment:
            for run in segment['transcriptSegmentRenderer']['snippet']['runs']:
                text = run['text']
                yield text.strip()

def main():
    # Read input JSON from stdin
    input_json = sys.stdin.read()
    data = json.loads(input_json)

    segments = get_segments(data)

    # Process the JSON and concatenate the transcript segments
    concatenated_string = ' '.join(extract_transcript_segments(segments))

    # Write the output to stdout
    sys.stdout.write(concatenated_string + '\n')

if __name__ == "__main__":
    main()

Usage

python youtube_transcript_to_md.py < transcript.json > transcript.md