Founder-operator. I build the enrichment, outbound, content and reporting systems that let small teams generate pipeline without proportional headcount — most of it by directing AI coding agents, with cost per outcome as the design constraint.
🔗 gtmstacker.com — the practice · Agent Registry — an agent-native registry of 300+ open-source GTM tools (llms.txt, OpenAPI, MCP server, JSON-LD; verified by 2,800+ automated checks) · LinkedIn · Writing
These repos weren't written in a weekend. They're internal builds from my practice (GTM Stacker, 2024–2026) — tooling that ran real client work. In September 2026 I started moving them into public: stripping client data and configuration, writing honest READMEs (including what each tool does not prove), and open-sourcing the ones that stand alone. More are queued as they get de-clientified.
Prospecting radars — find companies with a reason to buy (all on official Apify actors)
hiring-strain-radar ·
saas-displacement-finder ·
app-regression-radar ·
review-weakness-finder
Outbound & data infrastructure
email-list-scrubber — two-stage self-hosted
verification (email-verifier → reacher), no per-verification meter ·
plg-usage-signal-outreach — product
usage → sales outreach (PostHog → n8n → LiteLLM → Twenty), human-in-the-loop ·
youtube-transcript-harvester
Creative & content engines
adsmith — thousands of on-brand ad variants at $0/image
(deterministic templates + agentic SVGs) ·
slidesmith — token-driven, fully offline social-slide
renderer ·
x-viral-tracker — deterministic virality scoring
for niche X/Twitter monitoring
GEO / AI-search visibility
ai-subquery-inspector — observe the
web-search sub-queries AI assistants actually run, from your own browser session
Next out of the queue: page-auditor — grades a URL against a dated SEO/GEO rubric
(raw-vs-rendered content delta, JSON-LD, llms.txt) and states plainly what a URL fetch cannot
reveal. Currently being de-clientified; landing here shortly.
House rule across all of it: every README states limits plainly — what the tool does, and what it does not prove. If you're evaluating whether these patterns fit your stack, that's exactly what my intro call is for.