Automation

From YouTube to Blog Post: How I Built a Self-Maintaining Website

February 15, 2026

In this post I walk through the three automation pipelines that keep this website up to date without any manual intervention.

TL;DR

• A daily GitHub Actions cron job fetches my latest YouTube video
• Downloads the transcript using youtube-transcript-api
• Sends it to Claude (claude-sonnet-4-20250514) to generate a structured blog post
• Commits the Markdown file and triggers a Hugo rebuild

Key Takeaways

1. The whole process takes about 60 seconds from video publish to draft post
2. The Claude prompt is the most important part — iteration matters
3. Always diff before commit to avoid unnecessary rebuilds

The Pipeline

Step 1: Detect new video

import yaml, requests

def get_latest_video(api_key, channel_id):
    url = "https://www.googleapis.com/youtube/v3/search"
    params = {"key": api_key, "channelId": channel_id,
              "order": "date", "part": "snippet", "maxResults": 1}
    r = requests.get(url, params=params).json()
    return r["items"][0]

Step 2: Generate blog post with Claude

The Claude API call uses a structured prompt that ensures the output includes all required frontmatter fields and a consistent article structure.