obscura-fetch

name	obscura-fetch
description	Fetch a single web page using Obscura headless browser. Best for public, read-only pages. Returns HTML, plain text, links, or JS-evaluated results.

Fetch a URL with Obscura and return the content. Use this for single-page read operations — documentation, articles, public data pages.

Usage

/obscura-fetch <url> [--dump html|text|links] [--eval <js>] [--selector <css>] [--stealth]

Decision: which flag to use

Goal	Flag
Read article / docs	--dump text
Extract links	--dump links
Get raw HTML	--dump html (default)
Extract structured data	--eval "JSON.stringify(...)"
Wait for dynamic content	--selector <css> or --wait-until networkidle0
Bot-detection suspected	--stealth

Instructions

1. Parse URL and options from user input.
2. Run the command:

   obscura fetch <url> --quiet [options]
   

3. If output > 5000 chars, summarize key content — don't dump raw.
4. If exit code ≠ 0, report the error and suggest --stealth or --wait-until networkidle0.

Known limitations

• No session/cookie persistence across calls
• No click, form input, or navigation
• Login-required pages → not supported (suggest Playwright instead)

Adaptive extraction (for agentic loops)

When a fetch returns unexpected results in an automated workflow:

Symptom	Try
Empty body	--stealth
Bot detection page	--stealth --wait-until networkidle0
SPA content missing	--selector <main-element> --wait-until networkidle0
Redirect loop	Check final URL in output, fetch that directly
403 Forbidden	--stealth with --user-agent "Mozilla/5.0..."

URL canonicalization hint

Before adding a URL to a visited set, normalize it:

• Strip fragment: https://example.com/page#section → https://example.com/page
• Strip trailing slash: https://example.com/page/ → https://example.com/page
• Remove tracking params: utm_source, utm_medium, fbclid, gclid
• Sort query params alphabetically

Examples

# Read article as clean text
obscura fetch https://example.com --quiet --dump text

# Extract page title
obscura fetch https://example.com --quiet --eval "document.title"

# Extract all links
obscura fetch https://example.com --quiet --dump links

# Structured data extraction
obscura fetch https://news.ycombinator.com --quiet \
  --eval "JSON.stringify(Array.from(document.querySelectorAll('.titleline > a')).map(a => ({title: a.textContent, url: a.href})))"

# Wait for SPA content
obscura fetch https://example.com --quiet --selector "#main-content" --dump text

# Stealth mode for bot-protected pages
obscura fetch https://example.com --quiet --stealth --dump text