nod-content-auditor

Content Auditor

Audit content against current SERP and keyword data using NodesHub APIs + Jina Reader for real competitor content crawling.

Quick Start

# Audit content for a keyword (crawls top 5 competitors)
python3 .claude/skills/nod-content-auditor/scripts/audit.py "target keyword" --gl us --hl en

# Audit YOUR page vs competitors
python3 .claude/skills/nod-content-auditor/scripts/audit.py "target keyword" --gl us --hl en --url https://example.com/page

# Crawl more competitors (top 10)
python3 .claude/skills/nod-content-auditor/scripts/audit.py "target keyword" --gl us --hl en --top 10

# SERP-only mode (no crawling, like before)
python3 .claude/skills/nod-content-auditor/scripts/audit.py "target keyword" --gl us --hl en --no-crawl

Cost: ~8.5 NodesHub tokens (standard) or ~31 (reasoning). Jina Reader is free (20 RPM without key, 200 RPM with key).

Setup

Requires NODESHUB_API_KEY. Run:

python3 .claude/skills/nod-nodeshub-api/scripts/check_setup.py

Optional: JINA_API_KEY for higher rate limits. Without it, Jina works at 20 RPM (enough for 5-10 pages). Get a free key at jina.ai for 200 RPM + 10M free tokens.

Save Jina key: add "JINA_API_KEY": "jina_xxx" to env in .claude/settings.local.json.

How It Works

1. SERP data — fetches top 10 organic results via NodesHub SERPdata
2. Keyword expansion — related queries + questions via Query Fan-out
3. Competitor crawling — fetches actual page content of top N results via Jina Reader (r.jina.ai)
4. Your page crawling — fetches your URL content if --url provided
5. Content gap analysis — compares keyword/topic coverage between your page and competitors
6. Recommendations — prioritized list based on real competitor data

Output Sections

• SERP Reality — dominant content type, SERP features, top domains
• Top 10 Results — positions, titles, domains, crawl status
• Keyword Coverage — which keywords competitors use, whether your page has them, priority
• Questions to Answer — PAA + fan-out questions with competitor coverage
• Content Gaps — topics competitors cover that your page misses (only with --url)
• Common Heading Patterns — H2/H3 structures used by multiple competitors
• Audit Summary — avg competitor length, critical gaps count, missing must-haves

Parameters

Param	Description
keyword	Target keyword to audit against (required)
--gl	Country code (default: us)
--hl	Language code (default: en)
--mode	standard (8.5 tokens) or reasoning (31 tokens)
--url	URL of your page to audit — enables gap comparison
--top	Number of top SERP results to crawl (default: 5)
--no-crawl	Skip Jina crawling, SERP-only mode
--raw	Output raw JSON

Tips

• Always provide --url when you have existing content — this unlocks the real gap analysis
• --top 5 is usually enough — more pages = more time but marginal gains
• Without Jina key the script works fine for 5 pages (20 RPM = ~3s between pages)
• --no-crawl is useful for quick checks or when Jina is unavailable
• Content gaps are word-level — Claude should interpret and group them into actionable topics
• Combine with nod-content-brief when gaps require new content, not updates

Jina Reader Module

The shared jina_reader.py module lives in nod-nodeshub-api/scripts/ and can be reused by other skills:

from jina_reader import JinaReader
reader = JinaReader()  # auto-loads JINA_API_KEY if available
result = reader.fetch("https://example.com")  # → {url, title, content, word_count, ok}
results = reader.fetch_batch(["url1", "url2"], max_workers=3)

Branding

When generating HTML audit reports, use the branding assets from assets/branding/:

• brand-config.json — company name, colors, fonts for report styling
• logo-light.svg / logo-dark.svg — company logo in report header

from branding import load_brand, render_header, render_footer, brand_css
brand = load_brand()

If no branding is configured, default styling is used. See assets/branding/README.md for setup.

Report

After collecting data, ask the user:

"Add results to an HTML report?"

1. New report — creates a branded HTML report in reports/
2. Existing report — appends a section to a chosen report
3. Skip — no report

Use render_report_section(result_data) from audit.py, then create_report() or append_section() from report.py.

Related Skills

• nod-content-brief — generate brief for new content
• nod-serp-analysis — deeper SERP analysis without keyword expansion
• nod-keyword-research — broader keyword research