AI Citation Lead Extractor

See which domains Google AI Overview cites — turn AI sources into your lead list.

---

The Problem

SEO agencies, PR firms, and B2B content teams need to know which domains get cited for their target queries — but:

• Ahrefs/SEMrush cost $99–$299/mo with monthly commitments
• Raw Apify scrapers give you unstructured JSON — you must write the extraction logic
• Custom scraping requires engineering time, proxy management, anti-blocking

Meanwhile, Google AI Overview already exposes this data in its sources field. You just need it aggregated, deduped, and ready to import.

---

The Solution

Run a batch of prompts. Get back a deduped table of cited domains with URLs and prompt context.

What you get:

• A CSV/JSON table of cited domains with:
  • domain (deduped)
  • citation_count (how many prompts cited this domain)
  • sample_urls (actual URLs that were cited)
  • prompts_that_cited (which queries triggered the citation)

Example use case:

"Run 50 prompts about 'best AI chips'. Get back a clean list of 12 unique domains that AI Overview cited most — with URLs and prompt context. Import into CRM or outreach tool. Done."

---

Who Buys This (ICP)

Segment	Use Case
SEO Agencies	Link prospecting, competitor backlink analysis
PR Firms	See which outlets/brands get cited for topics
Content Teams	Find authoritative sources for their own content
B2B SaaS	Identify which competitors dominate AI answers for their keywords

---

How It Works

1. You provide a list of prompts (natural-language queries)
2. We call Apify's google-search-scraper with AI Overview add-on
3. We parse each result's aiOverview.sources array
4. We deduplicate by domain and aggregate citation counts
5. You get back a clean lead table (CSV/JSON)

Host Actor: apify/google-search-scraper (4.6★, 15K monthly users, actively maintained)

---

Input Schema

{
  "prompts": [
    "best AI chips for training large language models",
    "leading companies in generative AI infrastructure",
    "top SEO tools for link building in 2026"
  ],
  "country": "US",
  "language": "en",
  "min_citations": 1,
  "max_results": 100
}

Fields:

• prompts (required): List of 1–100 natural-language queries (max 500 chars each)
• country (optional): Google search country code (default: "US")
• language (optional): Google search language code (default: "en")
• min_citations (optional): Minimum citation count to include domain (default: 1)
• max_results (optional): Maximum unique domains to return (default: 100, max: 1000)

---

Sample Output

{
  "metadata": {
    "generated_at": "2026-08-24T12:34:56Z",
    "total_prompts": 50,
    "unique_domains": 12,
    "total_citations": 47,
    "data_mode": "live"
  },
  "leads": [
    {
      "domain": "nvidia.com",
      "citation_count": 12,
      "sample_urls": [
        "https://www.nvidia.com/en-us/data-center/h100/",
        "https://blogs.nvidia.com/blog/ai-chip-guide/"
      ],
      "prompts_that_cited": [
        "best AI chips for training",
        "leading companies in generative AI"
      ],
      "first_seen": "2026-08-24T12:34:56Z"
    },
    {
      "domain": "semrush.com",
      "citation_count": 8,
      "sample_urls": [
        "https://www.semrush.com/blog/seo-link-building/"
      ],
      "prompts_that_cited": [
        "top SEO tools for link building in 2026"
      ],
      "first_seen": "2026-08-24T12:34:56Z"
    }
  ]
}

---

Pricing

$0.018/prompt (target: 58% gross margin vs $0.0075 COGS)

Examples:

• 100 prompts = $1.80
• 500 prompts = $9.00
• 1,000 prompts = $18.00

Why this price?

• COGS: $0.0075/prompt (Apify's $0.0045/page + $0.003/AI Overview)
• We add value: aggregation, deduplication, lead table formatting
• Still 6–10x cheaper than Ahrefs/SEMrush subscriptions

---

Why Us vs. Incumbents

Feature	Raw Apify Scraper	Ahrefs/SEMrush	AI Citation Lead Extractor
Output format	Raw JSON (sources scattered)	Dashboard + CSV export	Deduped lead table (ready to import)
Commitment	Pay-per-use	$99+/mo subscription	Pay-per-use
Setup time	Write extraction code	Dashboard learning curve	Prompts → leads in minutes
AI Overview focus	Optional add-on	Limited AI coverage	AI sources only
Self-hostable	Yes	No	Yes
Price per 1k prompts	$7.50 (cost only)	~$100/mo unlimited	$18

---

Honest Limitations

What this is NOT:

• ❌ A full web scraper (we only extract AI Overview sources)
• ❌ A historical database (fresh extraction each run)
• ❌ Brand monitoring (that's our ../apify-actor product — checks if YOUR brand is cited)
• ❌ A monthly subscription (pay-per-use only)

Current constraints:

• Requires Apify billing unlock (see ops/LOGAN_UNLOCKS.md)
• Only supports Google AI Overview today (Perplexity, ChatGPT Search coming soon)
• maxTotalChargeUsd capped at $0.05 for smoke tests

PPE Events (not yet enabled):

• prompt_set_processed — one batch of prompts processed (primary event)
• Alternative models: per prompt ($0.015–0.025), per batch ($25–50), per output row ($0.10–0.20)

---

APIFY SETUP

Required:

1. Apify account with billing unlock (see ops/LOGAN_UNLOCKS.md)
2. APIFY_TOKEN environment variable set

Local testing:

cd actor2
export APIFY_TOKEN="your_token_here"
python3 main.py

Input file:

echo '{"prompts": ["best AI chips", "leading AI companies"]}' > input.json
python3 main.py

---

Output Formats

Results are pushed to Apify's default dataset in JSON format. You can:

• Export as CSV from Apify dashboard
• Download via Apify API
• Self-host and parse the JSON directly

---

FAQ

Q: How is this different from your AI Citation Monitor?
A: AI Citation Monitor checks if YOUR brand is cited. This extracts WHO is getting cited (competitor intel + lead gen).

Q: Can I export to CSV?
A: Yes — Apify's dashboard lets you export dataset results as CSV.

Q: What if Google changes the AI Overview structure?
A: We use selector-based parsing. If structure changes, we'll update the parser and notify users via Apify's changelog.

Q: Do you store historical data?
A: No — each run is a fresh extraction. We don't maintain a database of past results.

Q: Can I increase the $0.05 charge cap?
A: The cap is for smoke tests. Production runs will have higher limits (configured per user billing tier).

---

Related Products

• ../apify-actor — Check if Google AI Overview cites YOUR brand/domain
• Both products share the same COGS model ($0.0075/prompt) and billing infrastructure

---

Technical Details

Architecture:

• Host: apify/google-search-scraper (AI Overview add-on enabled)
• Worker: worker/core.py::extract_cited_leads()
• Language: Python 3.11+
• Framework: Apify SDK (async)

Dependencies:

• apify (Apify SDK)
• apify-client (Apify API client)

Self-hosting:

git clone https://github.com/log-wade/spark-agent-commerce.git
cd spark-agent-commerce/actor2
pip install -r requirements.txt
export APIFY_TOKEN="your_token"
python3 main.py

---

Support

• Email: [not yet configured]
• GitHub Issues: [not yet configured]

---

License

MIT — see LICENSE.txt in the repo root.

---

Last Updated: 2026-08-24
Version: 0.1.0 (MVP)
Owner: Spark Agent Commerce Team