Nothing here was built to be sold — it was built because my companies needed it, hardened in daily production, and then packaged. Buy a pack, install it into your Claude Code / Cursor workspace, and it works the same day. Every pack ships with docs and a walkthrough.
Everything powering the demo you just touched: build your own scored, agent-queryable entity graph for any domain, and put a touchable demo of it in front of your customers.
Point it at a Gmail account and it maps your real network: who you actually know, how warm every relationship is, who went quiet, and who can introduce you to anyone. Running on my own 306,000-message inbox right now.
The 15-role AI agent org my company runs on: org chart, autonomy rules, outbox/memo conventions, night-run scripts, onboarding docs. A three-person team shipping like twenty.
The quality layer for agent-built software: auth crawler, style linter, design-review agent, E2E test personas, overlap detector. The reason a solo-run agent org ships without breaking things.
The production skeleton behind six live inbox agents: router → rate limiter → abuse guard → intent → reply, with email and iMessage adapters. Build a customer-facing agent on a chassis that's already survived strangers.
The measurement layer from the cost-audit case study: event schema, per-user counters, cost dashboards, and the audit playbook. Know your cost-per-user-action before it eats your margin.
The full workflow that shipped a paying product in 72 hours: worktrees, smoke tests, deploy scripts, judge/eval patterns. The process, written down and runnable.
All seven packs, plus updates as they ship. If you're building a company on agents, this is the whole toolbox at half the à-la-carte price.
Checkout is Stripe; access lands in your inbox the same day, and every purchase includes a 30-minute install call. Need the outcome without the install? The services start at $15K.