Projects

Things I’m building or have built. Most are works in progress. Some are internal and won’t be public. I keep a long tail of small builds because that’s where most of my learning happens.

Strategy Labs

Fulcrum

The brand intelligence platform behind Strategy Labs. Seven years as an internal data warehouse, now being productized as standalone SaaS. Connects marketing, financial, and customer data into one ecosystem, then layers a growing set of AI tools on top — dashboards, benchmarking, call intelligence, agent-driven analysis. The big one.

Call intelligence platform

Its own project, but part of the Fulcrum ecosystem. Telephony-agnostic pipeline — ingests from CallRail or RingCentral via API, or accepts direct transcript uploads. Each recording gets transcribed, analyzed by Claude (type, sentiment, lead quality, issue categories, recommendations), embedded, and stored in PostgreSQL with pgvector. A Claude desktop companion lets any team member query the full history in plain English.

Strategy Labs Plugin Marketplace

Claude skill and plugin distribution for the agency’s internal team and select clients. Staff browse and install AI capabilities the way they’d install an app — no technical setup needed.

colab-benchmarking

Standalone proof-of-concept for the Fulcrum benchmarking module. Clients log in and see core ad performance against peer benchmarks.

AidKit

Meta Ads naming convention generator. React, TypeScript, Supabase. Small tool, constant pain point in ad ops.

Applied AI products

Food-cost intelligence platform (private beta)

Food cost intelligence for independent restaurant operators. A supplier invoice photo turns into structured line-item data — SKU, quantity, unit price — extracted via Claude vision, normalized across suppliers, stored for trend analysis. Price swings trigger alerts tied to specific menu items and recipe margins. In private beta with one design partner.

Stack: Claude vision, Trigger.dev, Next.js, Neon + pgvector, Clerk.

FencePro

Business management platform for fence contractors. Estimating, job tracking, and client management — with AI handling quote generation and follow-up drafts.

Agent and AI infrastructure

open-brain

MCP server that gives AI agents persistent semantic memory and RAG document search over a Supabase backend. The memory layer I plug into other projects. Open source.

Text provenance plugin (in progress)

Obsidian plugin for tracking which parts of a document were written by a person versus generated by AI. Provenance for when that distinction matters.

Smaller builds

  • mailtop — Terminal email client with AI triage, inspired by btop and mactop.
  • daylog — macOS menu-bar app that passively captures daily computer activity so I can ask agents “what did I actually work on?” with real data.
  • fireflies-obsidian — Auto-imports Fireflies meeting transcripts into my Obsidian vault.
  • clief-notes — Constraint library, reference architecture gallery, and skill starter for building with Claude Code.

A note on quality

Some of these are polished. Many are scrappy. A few are in-progress. Read each one as a thing I wanted to learn or solve at the time.