What problem does it solve?
Cold email campaigns often need to know whether each prospect's own website mentions a specific thing (a certification, technology, or service) when no structured data provider sells that field, and turning that signal into safe, non-fabricated personalization copy is error-prone.
Core Features & Use Cases
- Keyword presence detection: Runs one literal Google
site:domain "keyword" query per row through a SERP API, with a deterministic literal-phrase filter and subdomain exclusions before any model call.
- Gated copy generation: A nano-class model judge classifies each hit (has/planned/discusses/none) with confidence, and only
has + high rows produce a copy-ready site_keyword_sentence like "Noticed you hold SOC 2 Type II certification."
- Dual implementation: Ships as both a Clay table build (11 columns with dependency-cycle handling) and a Clay CLI workflow (6 nodes), with measured 90% accuracy at roughly $0.59 per 1,000 rows.
- Use Case: Given a list of 500 B2B domains, filter to companies whose sites mention "SOC 2" and upload only the gated, pre-rendered personalization sentences to your email sequencer.
Quick Start
Ask Claude to check which companies in my lead list mention SOC 2 on their website using the playbook-google-site-search skill and generate a personalization line for each hit.
