gkwa YouTube Ingest Workflow
Given a YouTube URL (e.g. https://youtu.be/0EHFgu7V7N4), run the following steps in order.
Carry outputs forward — do not prompt the user between steps unless a step fails.
Step 1: Fetch video metadata
Run yt-dlp to get the video title:
uvx yt-dlp --get-title "<youtube_url>"
Collect from the output:
- The video title
Also fetch the page to identify the channel name and a one-sentence summary of the topic. Use WebFetch on the canonical YouTube watch URL with a prompt asking for title, channel, and topic.
Step 2: Take the thumbnail URL directly
YouTube already hosts the thumbnail at a stable URL, so do not run the social-to-imgur skill here.
Build the URL from the video id: https://i.ytimg.com/vi/<VIDEO_ID>/maxresdefault.jpg
Use that for both pic: and the hero image.
The Imgur round trip adds nothing for YouTube and runs into the egress-IP 429. Instagram and other platforms with expiring CDN URLs still need the upload.
Step 3: Derive the vault filename
Convert the video title to a vault-safe filename:
- All lowercase
- Words separated by spaces (no hyphens or underscores)
- Strip special characters (pipes, colons, slashes, etc.)
- Strip filler words that add no meaning (e.g. "how to", "a", "the") only when the result is still descriptive
Example: "How To Make Dumpling Wrappers From Scratch | KitchenAid Pasta Attachment"
becomes: dumpling wrappers from scratch kitchenaid pasta attachment.md
Step 4: Write the vault note
Write the note to /Users/mtm/Documents/Obsidian Vault/<filename>.md with this structure:
<youtube_url>
No frontmatter is needed unless the user asks for it.
Step 5: Commit
Stage and commit the new file:
git add "<filename>.md"
git commit -m "Add <short topic> note with YouTube link and hero image"
Run each git command as a separate Bash call (no chaining with &&).
Final report
After all steps complete, report:
- The vault note filename
- The i.ytimg.com thumbnail URL
- The original YouTube URL