CSS integrity
StyleSafe
Catch CSS cascade conflicts, specificity problems, and clashing Tailwind utilities before they ship.
125/mo
PORTFOLIO / BUILT IN THE OPEN
Useful tools.
Unique builds.
Developer tools, independent products, and experiments that turn curiosity into something you can use. A look inside my workshop.
OPEN SOURCE / MCPStyleSafeCatch CSS cascade conflicts, specificity problems, and clashing Tailwind utilities before they ship.Explore this project OPEN SOURCE / MCPStyleSpeakResolve the cascade, trace a property, and preview the reach of a change before touching the stylesheet.Explore this project OPEN SOURCE / MCPEnvSpeakUnderstand which environment value wins across .env files, Docker Compose, and Kubernetes—and why.Explore this project 
SELECTED BUILDRegulatonAn AI governance product exploration for small businesses. From an inventory of AI tools to document workflows, compliance dashboards, and the infrastructure around a multi-tenant SaaS.Explore this project 
SELECTED BUILDNJCSVAn independent explorer for New Jersey Division of Family Development data. Follow program trends, compare counties, and connect the charts to the original published reports.Explore this project 
SELECTED BUILDThe Argument.aiA shared argument map where a person and an AI agent can think together. Claims, evidence, and objections become a visible graph that both sides can work on through WebMCP.Explore this project THE WORKBENCHPhonetic WorkbenchChoose sounds letter by letter, refine a pronunciation, and export respellings, Arpabet, or ElevenLabs phoneme tags. The workbench behind Miss Pronunciation.Explore this project THE WORKBENCHAgent PrimitivesSmall, composable instruction files for better coding-agent behavior: scope, verification, codebase alignment, and the habits that make collaboration work.Explore this project THE WORKBENCHEnterprise RAG · NeonA TypeScript reference backend combining local embeddings, tenant and role filters, source citations, and a separate answer-verification pass.Explore this project THE WORKBENCHCodebase Consistency HooksA pattern-convergence checker that finds established codebase conventions and flags new changes that diverge from them. One Python engine, thin adapters.Explore this project THE WORKBENCHGuess My WeightsA metacognitive chat experiment that surfaces the weighted emphasis of expressed beliefs, preferences, and values so you can examine what is shaping the conversation.Explore this project
SELECTED BUILDRegulatonAn AI governance product exploration for small businesses. From an inventory of AI tools to document workflows, compliance dashboards, and the infrastructure around a multi-tenant SaaS.Explore this project SELECTED BUILDNJCSVAn independent explorer for New Jersey Division of Family Development data. Follow program trends, compare counties, and connect the charts to the original published reports.Explore this project SELECTED BUILDThe Argument.aiA shared argument map where a person and an AI agent can think together. Claims, evidence, and objections become a visible graph that both sides can work on through WebMCP.Explore this project THE WORKBENCHPhonetic WorkbenchChoose sounds letter by letter, refine a pronunciation, and export respellings, Arpabet, or ElevenLabs phoneme tags. The workbench behind Miss Pronunciation.Explore this project THE WORKBENCHAgent PrimitivesSmall, composable instruction files for better coding-agent behavior: scope, verification, codebase alignment, and the habits that make collaboration work.Explore this project THE WORKBENCHEnterprise RAG · NeonA TypeScript reference backend combining local embeddings, tenant and role filters, source citations, and a separate answer-verification pass.Explore this project THE WORKBENCHCodebase Consistency HooksA pattern-convergence checker that finds established codebase conventions and flags new changes that diverge from them. One Python engine, thin adapters.Explore this project THE WORKBENCHGuess My WeightsA metacognitive chat experiment that surfaces the weighted emphasis of expressed beliefs, preferences, and values so you can examine what is shaping the conversation.Explore this project
01 / OPEN SOURCE
A better way
to develop.
Cascade insight
StyleSpeak
Resolve the cascade, trace a property, and preview the reach of a change before touching the stylesheet.
403/mo
Configuration clarity
EnvSpeak
Understand which environment value wins across .env files, Docker Compose, and Kubernetes—and why.
134/mo
02 / SELECTED BUILDS
From a simple idea
to a working creation.
BUILD.01AI governance / SaaS
regulaton.vercel.appExplore the build AI governance / SaaS
Regulaton
An AI governance product exploration for small businesses. From an inventory of AI tools to document workflows, compliance dashboards, and the infrastructure around a multi-tenant SaaS.
- AI tool inventory and document workflows
- Authentication, subscriptions, and webhooks
- A complete product built around a real workflow
BUILD.02Public data / visualization
nj-csv.vercel.appExplore the build Public data / visualization
NJCSV
An independent explorer for New Jersey Division of Family Development data. Follow program trends, compare counties, and connect the charts to the original published reports.
- Source-linked trends and county comparisons
- CSV exports and transparent methods
- Forecasting experiments with time-based backtesting
BUILD.03Human + agent / WebMCP
the-argumentai.vercel.appExplore the build Human + agent / WebMCP
The Argument.ai
A shared argument map where a person and an AI agent can think together. Claims, evidence, and objections become a visible graph that both sides can work on through WebMCP.
- A shared visual argument graph
- Agent tools for claims, objections, and rebuttals
- Browser-local persistence with no backend
03 / THE WORKBENCH
Follow the
interesting problem.
Voice / language
Phonetic Workbench
Choose sounds letter by letter, refine a pronunciation, and export respellings, Arpabet, or ElevenLabs phoneme tags. The workbench behind Miss Pronunciation.
Agent workflows
Agent Primitives
Small, composable instruction files for better coding-agent behavior: scope, verification, codebase alignment, and the habits that make collaboration work.
Retrieval / reference backend
Enterprise RAG · Neon
A TypeScript reference backend combining local embeddings, tenant and role filters, source citations, and a separate answer-verification pass.
Developer tooling
Codebase Consistency Hooks
A pattern-convergence checker that finds established codebase conventions and flags new changes that diverge from them. One Python engine, thin adapters.
Conversational AI / experiment
Guess My Weights
A metacognitive chat experiment that surfaces the weighted emphasis of expressed beliefs, preferences, and values so you can examine what is shaping the conversation.
There is always
another idea.
The rest of the workshop lives on GitHub.
Browse all repositories (opens in a new tab)NEXT / A GOOD CONVERSATION
Something worth
building together?
Let’s talk about the work, the problem, and what comes next.
