Shipped work

Projects

Studio work, client sites, and independent products. Explore the work and open a case study for the details.

Showing 5 Product projects

Product2026Live

FreeVoiceTranscribe

Private, local hold-to-talk dictation for Apple Silicon Macs. Hold fn, speak, then release to insert. On-device Whisper, no cloud accounts, no always-on mic.

Role: Solo product engineer (Python / macOS)
Outcome: Open-sourced an MIT-licensed menu-bar dictation app with local inference, hands-free mode, tests, and packaging scripts.

PythonWhispermacOSMLX

View on GitHub

Case study

Challenge, approach, and impact for FreeVoiceTranscribe

Challenge: Make fast dictation private and practical for Apple Silicon users without a cloud account or always-on microphone.
Approach: Paired a hold-to-talk workflow with on-device Whisper inference, tests, packaging, and a native menu-bar experience.
Impact: Released an MIT-licensed local dictation app that keeps speech processing on-device.

Product2026Live

Apple App Intents Skill

A free, open-source agent skill for practical Siri and Shortcuts integration. Guides Codex, Claude Code, and OpenClaw through App Intents, entities, indexing, and testing, with a FieldNotes Swift example.

Role: Creator, researcher, and Swift example author
Outcome: Published an MIT-licensed SKILL.md and focused references, free to use, modify, and redistribute.

SwiftApp IntentsSwiftUIAgent Skills

Get the skill

Case study

Challenge, approach, and impact for Apple App Intents Skill

Challenge: Developers need to map app actions to Apple's supported Siri and App Schema contracts while keeping existing Shortcuts useful.
Approach: I researched Apple's documentation and WWDC26 sessions, then built a portable agent skill with schema guidance, entity queries, indexing, testing workflows, and a persistent FieldNotes SwiftUI example.
Impact: The public MIT-licensed skill gives teams reusable integration guidance. Siri availability depends on supported schemas, SDKs, and eligible devices; build checks alone do not verify a live Siri experience.

Product2026Prototype

Traces

Local-first desktop knowledge workspace: markdown vault, 3D force graph of wiki-links, and multi-provider AI chat that can read and edit notes.

Role: Founder product and desktop engineer
Outcome: Built an Electron + Next.js app with CodeMirror editing, React Three Fiber graph views, and vault-aware AI tooling.

ElectronNext.jsR3FTypeScript

View product site

Case study

Challenge, approach, and impact for Traces

Challenge: Keep a personal knowledge workspace on the machine, and still make note links easy to follow and edit.
Approach: I combined a markdown vault, a graph view, rich editing, and vault-aware AI tools in a desktop app.
Impact: The prototype lets you navigate and edit connected notes without sending the vault to a host.

Product2026Prototype

WealthWise

Next.js money dashboard. Import a CSV or PDF statement, chart it, or load the bundled sample. You run the app. No Plaid. No live bank login.

Role: Founder product and full-stack engineer
Outcome: Working prototype with CSV/PDF import, charts, sample data, and Supabase email login.

Next.jsTypeScriptSupabaseRecharts

View product site

Case study

Challenge, approach, and impact for WealthWise

Challenge: Review a bank statement on your own machine without handing over a live bank login.
Approach: I built a Next.js dashboard that imports CSV or PDF statements, charts them with Recharts, and ships a bundled sample. Login is optional Supabase email auth.
Impact: A public Pages site shows the prototype. The numbers there are the bundled sample CSV, not a live balance.

Product2026Live

Command.AI

Veteran-focused product for claim prep, benefits education, and financial readiness. Mission-style planning instead of scattered checklists.

Role: Founder and full-stack product engineer
Outcome: Shipped a live web app with auth, structured learning paths, and ongoing product iteration for service members in transition.

Next.jsTypeScriptConvex

Open platform

Case study

Challenge, approach, and impact for Command.AI

Challenge: Claims, benefits, and money prep for service members lived in scattered checklists.
Approach: I built an authenticated product with learning paths and mission-style next steps.
Impact: Service members can work a live plan instead of hunting through disconnected lists.