Web Fetch

You can fetch and read the content of any web page, API endpoint, or file URL.

Fetching web pages (readable text)

# Fetch page content — strip HTML, get readable text
curl -sL "URL" | sed 's/<script[^>]*>.*<\/script>//g' | sed 's/<style[^>]*>.*<\/style>//g' | sed 's/<[^>]*>//g' | sed '/^[[:space:]]*$/d' | head -300

# Using readability-cli if installed (much better extraction)
readable "URL" 2>/dev/null || curl -sL "URL" | sed 's/<[^>]*>//g' | sed '/^$/d' | head -300

# Fetch only the <article> or <main> content (more targeted)
curl -sL "URL" | grep -oP '<(article|main)[^>]*>.*?</(article|main)>' | sed 's/<[^>]*>//g' | head -200

Fetching JSON APIs

# Fetch and parse JSON
curl -s "API_URL" -H "Accept: application/json" | jq '.'

# Extract specific fields
curl -s "API_URL" | jq '.data[] | {id, name, value}'

# POST request with JSON body
curl -s -X POST "API_URL" -H "Content-Type: application/json" -d '{"key": "value"}' | jq '.'

Checking headers only (HEAD request)

# Check content type, size, redirects without downloading
curl -sI "URL" | head -15

Downloading files

# Download to a specific path
curl -sLo /tmp/output_file "URL"

# Download and check file type
curl -sL "URL" -o /tmp/downloaded && file /tmp/downloaded

Tips

• Always use -sL (silent + follow redirects).
• For large pages, pipe through head -N to limit output.
• Strip HTML with sed 's/<[^>]*>//g' for plain text.
• Check Content-Type header first to decide parsing strategy.
• Respect robots.txt and rate limits.
• Combine with web-search skill: search first, then fetch top results.
• For paywalled or JS-heavy sites, results may be limited.

Triggers

fetch, read this URL, open this link, what does this page say, get content from, scrape, ler esta página, abrir link