In my previous post, I documented a two-part system for automating quote post creation: a Node.js script and a Feedly-specific bookmarklet. Since then, I’ve extended the tooling to be more general-purpose and robust.

The Problem with Feedly-Only Extraction

The original bookmarklet relied on Feedly’s specific DOM structure, which meant it only worked when reading articles within Feedly’s interface. This was limiting—I wanted to capture quotes from:

• Articles linked in tweets
• Blog posts I encounter through search
• News sites I visit directly
• Any webpage with proper meta tags

The Solution: Universal Meta Tag Extraction

Instead of parsing Feedly’s DOM, I refactored the bookmarklet to extract metadata using standard HTML meta tags that most modern websites include:

1
2
3
4

<meta property="og:author" content="Author Name">
<meta property="og:title" content="Article Title">
<meta property="article:published_time" content="2026-02-07">
<meta property="article:tag" content="Technology">

What Gets Extracted Now

The bookmarklet automatically captures:

1. Author - With fallback chain:

   • article:author meta tag
   • og:author (Open Graph)
   • Site name from og:site_name or application-name
   • Falls back to “Unknown” if not found

2. Title - Tries multiple sources:

   • Open Graph og:title
   • Twitter Card twitter:title
   • HTML <title> tag as last resort

3. Publication Date - Extracts and validates ISO format (YYYY-MM-DD):

   • article:published_time
   • datePublished meta tag
   • Custom publish_date meta tag

4. Tags - Extracts all article:tag meta tags:

   • Can capture multiple tags from a single page
   • Pre-fills the Tags field in the form
   • Fully editable before generating the command

5. Highlighted Text - Uses window.getSelection() to capture:

   • Text you highlight/select BEFORE clicking the bookmark
   • Preserves your selection exactly as you made it

Auto-Fetch Metadata from URL

Both the bookmarklet and create-quote-post.js now also intelligently fetch missing metadata:

1
2
3
4
5

# If author is missing or "Unknown", the script fetches it from the URL
node tools/create-quote-post.js \
  --url "https://example.com/article" \
  --quote "The quote..." \
  --commentary "My thoughts..."

The script will:

• Fetch the article’s HTML
• Extract author, title, date, and tags from meta tags
• Log what it found: ✓ Found author: Ben Thompson
• Use the extracted values for post generation

The general bookmarklet works on any website with proper meta tags, pawing the path for future automations.

The updated resources are available at:

• feedly-bookmarklet.js
• general-bookmarklet.js
• create-quote-post.js