pratos/obsidian-link-archiver

Obsidian Link Archiver

Activation

When this skill is triggered, ALWAYS display this banner first:

╭─────────────────────────────────────────────────────────────╮
│  📥 SKILL ACTIVATED: obsidian-link-archiver                 │
├─────────────────────────────────────────────────────────────┤
│  Action: Extract links → Fetch content → Save locally       │
│  Output: Archived markdown files in same directory          │
╰─────────────────────────────────────────────────────────────╯

When to Use

• "archive all links in this note"
• "save linked content locally"
• "download all URLs from this markdown file"
• "backup external links in my obsidian note"
• "fetch and save all links from [file.md]"

Process

Step 1: Extract Links from Markdown File

Read the specified markdown file and extract all external links:

# Extract markdown links [text](url) and bare URLs
grep -oE '\[([^\]]+)\]\((https?://[^)]+)\)|(https?://[^\s\)]+)' "$FILE" | \
  sed -E 's/\[([^\]]+)\]\((https?:\/\/[^)]+)\)/\2/g' | \
  sort -u

Link types to extract:

• Markdown links: [Link Text](https://example.com)
• Bare URLs: https://example.com
• Reference links: [text][ref] with [ref]: https://...

Skip these:

• Internal Obsidian links: [[Note Name]]
• Anchor links: #section
• Local file links: file://...
• Image links (optional, based on user preference)

Step 2: Fetch Content via markdown.new

For each URL, fetch content as clean markdown:

# Fetch URL as markdown (free, no API key needed)
curl -sL "https://markdown.new/$URL" > "$OUTPUT_FILE"

Benefits of markdown.new:

• Converts HTML to clean markdown
• Strips ads, navigation, boilerplate
• Preserves content structure (headers, lists, code blocks)
• 80% fewer tokens than raw HTML

Step 3: Save Locally

Save each fetched page in the same directory as the source file:

# Directory of the source file
DIR=$(dirname "$SOURCE_FILE")

# Create archived-links subdirectory
mkdir -p "$DIR/archived-links"

# Save with sanitized title as filename
# Example: "My Article Title.md"

Filename format:

• Use page title from content (first # heading or <title>)
• Sanitize: remove special chars, limit length
• Add .md extension
• Handle duplicates: title.md, title-2.md, etc.

Step 4: Update Source File (Optional)

Optionally update the original file to reference local copies:

<!-- Original -->
[Article](https://example.com/article)

<!-- Updated -->
[Article](./archived-links/Article.md) ([original](https://example.com/article))

Example Workflow

Given a file notes/research.md:

# Research Notes

Found some great resources:
- [Stagehand Docs](https://docs.stagehand.dev/introduction)
- [Cloudflare Browser Rendering](https://developers.cloudflare.com/browser-rendering/)

Also check https://example.com/article

Run the archiver:

# 1. Extract links
links=(
  "https://docs.stagehand.dev/introduction"
  "https://developers.cloudflare.com/browser-rendering/"
  "https://example.com/article"
)

# 2. Create output directory
mkdir -p "notes/archived-links"

# 3. Fetch and save each link
for url in "${links[@]}"; do
  # Get page title for filename
  content=$(curl -sL "https://markdown.new/$url")
  title=$(echo "$content" | grep -m1 "^# " | sed 's/^# //' | tr -cd '[:alnum:] -' | cut -c1-50)
  
  # Fallback to URL-based name if no title
  if [ -z "$title" ]; then
    title=$(echo "$url" | sed 's|https://||; s|/|-|g' | cut -c1-50)
  fi
  
  # Save content
  echo "$content" > "notes/archived-links/${title}.md"
  echo "✅ Saved: ${title}.md"
done

Result:

notes/
├── research.md
└── archived-links/
    ├── Introduction - Stagehand Docs.md
    ├── Browser Rendering - Cloudflare Docs.md
    └── example-com-article.md

Complete Script

#!/usr/bin/env bash
set -euo pipefail

SOURCE_FILE="${1:?Usage: $0 <markdown-file>}"

if [ ! -f "$SOURCE_FILE" ]; then
  echo "❌ File not found: $SOURCE_FILE"
  exit 1
fi

DIR=$(dirname "$SOURCE_FILE")
ARCHIVE_DIR="$DIR/archived-links"
mkdir -p "$ARCHIVE_DIR"

echo "╭─────────────────────────────────────────────────────────────╮"
echo "│  📥 Archiving links from: $(basename "$SOURCE_FILE")"
echo "╰─────────────────────────────────────────────────────────────╯"

# Extract all URLs
urls=$(grep -oE 'https?://[^\s\)\]\"]+' "$SOURCE_FILE" | sort -u)

if [ -z "$urls" ]; then
  echo "No external links found."
  exit 0
fi

count=0
total=$(echo "$urls" | wc -l | tr -d ' ')

while IFS= read -r url; do
  ((count++))
  echo ""
  echo "[$count/$total] Fetching: $url"
  
  # Fetch as markdown
  content=$(curl -sL --max-time 30 "https://markdown.new/$url" 2>/dev/null || echo "")
  
  if [ -z "$content" ] || echo "$content" | grep -q '"success":false'; then
    echo "  ⚠️  Failed to fetch, trying direct curl..."
    content=$(curl -sL --max-time 30 "$url" 2>/dev/null | head -c 100000 || echo "")
    if [ -z "$content" ]; then
      echo "  ❌ Skipped (fetch failed)"
      continue
    fi
  fi
  
  # Extract title from content
  title=$(echo "$content" | grep -m1 "^# " | sed 's/^# //' | head -c 60 || echo "")
  
  # Fallback to URL-based name
  if [ -z "$title" ]; then
    title=$(echo "$url" | sed 's|https\?://||; s|[/?#]|-|g; s|-\+|-|g; s|-$||' | head -c 50)
  fi
  
  # Sanitize filename
  filename=$(echo "$title" | tr -cd '[:alnum:] ._-' | sed 's/  */ /g; s/^ //; s/ $//')
  
  # Handle duplicates
  output_file="$ARCHIVE_DIR/${filename}.md"
  if [ -f "$output_file" ]; then
    i=2
    while [ -f "$ARCHIVE_DIR/${filename}-${i}.md" ]; do
      ((i++))
    done
    output_file="$ARCHIVE_DIR/${filename}-${i}.md"
  fi
  
  # Add source URL header and save
  {
    echo "---"
    echo "source: $url"
    echo "archived: $(date -Iseconds)"
    echo "---"
    echo ""
    echo "$content"
  } > "$output_file"
  
  echo "  ✅ Saved: $(basename "$output_file")"
  
  # Rate limit
  sleep 0.5
  
done <<< "$urls"

echo ""
echo "╭─────────────────────────────────────────────────────────────╮"
echo "│  ✅ Done! Archived $count links to: archived-links/          "
echo "╰─────────────────────────────────────────────────────────────╯"

Parameters

| Parameter | Description | Default | |-----------|-------------|---------| | Source file | Path to the Obsidian markdown file | Required | | Output directory | Where to save archived content | ./archived-links/ | | Include images | Also download linked images | false | | Update source | Update original file with local links | false | | Max depth | Recursive link following depth | 1 (no recursion) |

Error Handling

• Timeout: Skip URLs that take >30s to fetch
• Failed fetch: Log warning, continue with next URL
• Duplicate title: Append number suffix (-2, -3, etc.)
• Invalid URL: Skip non-http(s) URLs
• Empty content: Skip and log warning

Limitations

• Only fetches first level of links (no recursion)
• Some sites may block markdown.new - falls back to direct curl
• Very large pages truncated to 100KB
• Rate limited to ~2 requests/second to be respectful

Integration with Obsidian

The archived files include YAML frontmatter compatible with Obsidian:

---
source: https://original-url.com
archived: 2026-03-04T17:00:00+05:30
---

This allows you to:

• Track where content came from
• Know when it was archived
• Use Obsidian's backlinks to find references