Local Business Opportunity Finder

Use case

Find evidence-supported local-business opportunities from public business rows you already have. Choose web_design, local_seo, or booking_automation. The Actor produces sorted prospect records, bounded page evidence, evidence IDs, and editable drafts. It never sends outreach or uses an LLM.

Buyer outcomes

• Turn an existing business export or Apify dataset into ranked service opportunities.
• Keep every finding tied to the exact public pages and retrieval time that support it.
• Separate verified opportunities, clean no_signal results, and unavailable unknown results instead of charging for silent empties.
• Get an agent-ready Markdown briefing plus structured dataset rows without providing an external AI key.

Example

{
  "mode": "imported_businesses",
  "businessType": "locksmith",
  "location": "Richmond, Virginia",
  "service": "local_seo",
  "maxBusinesses": 25,
  "importedBusinesses": [{ "businessName": "Example Locksmith", "websiteUrl": "https://example.com", "address": "Example address" }]
}

Use importedDatasetId instead of importedBusinesses when you have an Apify dataset the run token can read. It is paginated and only public-business fields are retained.

Output

The dataset writes sorted prospect records first and the final summary last. outputCounts reports the rows actually emitted as opportunity, no_signal, and unknown. A prospect includes bounded website evidence and a priority score.

• unknown has no asserted website-review conclusion, offer, or outreach draft.
• no_signal has no offer or outreach draft.
• web_design can say only listing has no website URL when that field is absent. It does not claim the business has no website.
• local_seo identifies missing title, metadata, and noindex only per checked page. A signal is not a guaranteed defect.
• booking_automation treats a missing anchor as a weak, limited observation, never proof that no booking process exists.

The Actor retains generic public business-email patterns and public business-phone values only. It does not retain reviewer identities, profiles, images, personal-contact enrichment, or send contact.

Charging

Each Verified business audit costs $0.005 through the successful-prospect-output event: supported opportunity rows and clean no_signal rows after a successful public website inspection. Buyers pay for verified filtering as well as supported findings. unknown, source failures, and listing-only rows with no independently inspected website remain nonchargeable. Buyers cannot enable or disable charging through input. Before each chargeable audit is published, the SDK checks that one whole event remains within the runtime charge limit and then writes it through Actor.pushData(row, 'successful-prospect-output'); no separate precharge occurs. The final dataset summary is written last and reports the actual emitted and charged verified-audit counts. Local non-monetized runs remain usable.

Maps dependency

maps_discovery calls compass/crawler-google-places with enrichment, images, and review-personal-data extraction disabled. It requires an Apify plan that permits the public Actor dependency. If that dependency is unavailable or not SUCCEEDED, the run fails clearly and does not return an empty successful result. The upstream Maps Actor may charge separately under its own pricing. Those charges are not included or measured here. Imported-business and imported-dataset modes do not require that upstream dependency.

FAQ

Does it contact businesses? No. Drafts are editable and remain unsent.

How much of a website is checked? Up to three same-origin public HTML pages. Credentials, custom ports, private/reserved targets, redirects to unsafe targets, and overlarge responses are rejected. Production requests use DNS-pinned transport to prevent rebinding.

Can I use a Maps dataset I already own? Yes, provide importedDatasetId with a token authorized for that dataset.

Support

Include the run ID, input mode, and the evidence IDs or URLs you want reviewed. Do not include credentials or personal-contact data.

Development

The image uses Node 22.

npm ci
npm run typecheck
npm test
npm run build
npm run smoke
docker build -t local-business-opportunity-finder .