Home  /  Workshop
The bench · The log · The shelf

The workshop.

This is the back room — what’s actually on the bench right now, what’s parked on the shelf for later, and the running notes I keep while figuring it out. The Studio page is the polished work. This is the real one.

5active builds
4on the shelf
30log entries
Updated weekly
01 · On the bench

What I’m actually building right now.

As of May 04, 2026Honest count, not a roadmap.
01 · ActiveShipping

League-OS 1.2.

Custom league management for YLAFL — registration, scheduling, comms, payments. Live with 1,500 families and still the most active codebase I own. New features land almost every week; the admin surface keeps getting tighter while the operator load keeps getting lighter.

What I’m chewing on

How to keep the admin UI brutally simple as the feature surface grows — and what stays YLAFL-specific vs. what generalizes for the next league. Every feature gets a “would Paul use this on a Saturday” test.

Last touched May 01w/ Paul Albertini
02 · ActiveBuilding

Roof Guru.

Roofing estimator and job platform for a Nashville operator. Intake from HOVER PDFs, cascading material templates, labor + pricing grids, commission projection. Parity with the spreadsheet first, then everything the spreadsheet couldn’t do.

What I’m chewing on

Every roofer has a spreadsheet that works. The only way to earn the switch is to match their numbers exactly on day one — then make the next estimate take five minutes instead of forty-five.

Last touched Apr 27w/ Ron
03 · ActiveBuilding

Jettison Air.

Aviation-adjacent operations platform built with Cole. Trip management, profit math, audit log, a public API, and an admin surface that treats the operator’s time as the scarcest resource. Replacing a spreadsheet one feature at a time.

What I’m chewing on

Where to draw the line between what the platform automates and what the operator still wants to touch. Too much automation loses trust; too little and we’re just a prettier Excel.

Last touched Apr 30w/ Cole Hambright
04 · ActiveResearching

Personal Health Engine.

A private dashboard that pulls labs, lifts, sleep, and food into one honest picture — and tells me what to actually change next, not what’s “trending.” Research mode for now: collecting signal, sketching rules, trying things on myself before anyone else.

What I’m chewing on

How much of the “what to change” should be model-generated vs. rule-based. Models are smarter, rules are accountable. Probably both, with rules in the loop.

Last touched Apr 17Solo research
05 · ActivePaused

Biblical Battle Plans.

A study tool that maps narrative arcs and recurring patterns across scripture — built from the way I personally read and want to study. The one I want to be working on most.

What I’m chewing on

Deprioritized against products with nearer-term revenue. Not shelved — just waiting for the right opening. When the billable work has a clear lane, this is the first thing I come back to.

Last touched Apr 08Solo build · paused
02 · Field log

Notes from the bench.

30 entriesMost recent first
Jun 242026
Wednesday
13:00
Build

League-OS: parents get an inbox.

Until now the league could only reach parents by email blast, with no record a family could go back and re-read. League-OS adds an in-app inbox: admins post an announcement, the targeted families see it in the parent portal with an unread badge, and it stays there to open again. League news finally has a home parents can return to, not just an email that scrolls away.

Jun 182026
Thursday
13:34
Build

League-OS: best fit for the open spot.

When a player drops a team, the waitlist admin can now pull the best replacement for that specific team instead of just the next name in line. League-OS scores each waitlisted player against the open team on the signals that matter, requested friends, same school, how close they live, and do-not-pair conflicts, then surfaces them as plain great, good, or okay tiers with conflicts flagged red. Filling a roster spot becomes a judgment call the software actually helps with.

Jun 182026
Thursday
00:07
Build

Biblical Battle Plans: a front door for the adventure.

Biblical Battle Plans replaced its bare landing page with a full long-scroll marketing page in the app's pixel-art RPG style: a parchment hero, a feature grid, an animated progression ladder, a product showcase, and a guild spotlight, with sprites and parallax that move as you scroll. It's the first surface that sells the app the way the app actually feels to play. The front door finally looks like the adventure waiting inside it.

Jun 172026
Wednesday
19:44
Build

Cornerbell: a front door of its own.

Cornerbell now has a front door of its own: a product-marketing site on the base domain with home, features, and about pages, capped by a contact page whose demo-request form and waitlist feed a single lead endpoint. The platform that hosts everyone else's site can finally pitch and capture interest in itself. The thing that builds other people's sites now has one worth visiting.

Jun 132026
Saturday
14:16
Build

Stunt Dudes: off Wix, onto its own stage.

Stunt Dudes traded a Wix template for a custom site: a dark tour-poster build with shows, a media gallery, a tour schedule, and a booking-inquiry form, all editable through a real admin CMS instead of a page builder. Structured data, an llms feed, and per-page metadata went in from day one so promoters and search engines can both find them. A stunt team's web presence that finally hits as hard as the act.

Jun 102026
Wednesday
16:22
Build

Cornerbell: one platform, many tenants.

Cornerbell crossed from being one promotion's website to a platform any organization can stand up as its own. A fresh tenant now provisions in a single command, brands itself, and runs its own teams, partners, photos, and media from the dashboard, with every public page served from that tenant's content behind a private, per-org media pipeline. Sign on, and the site is entirely yours, not a skin over someone else's.

Showing 6 of 30
03 · The shelf

What I’m not building.

4 ideasParked, not killed
01Parked

Builder’s Ledger.

A personal ledger app for solo founders — not finance, but decisions. What you committed to, what you said no to, what you actually shipped, audited monthly.

Why parkedI’d use it. Three other people I trust would use it. That’s not a market — that’s a Notion template.
02Parked

K2500 Build Log.

A small public site documenting the square-body restoration — parts, sources, what broke, what worked. Not a business; a useful artifact for the next guy.

Why parkedWill probably ship eventually. Static site, low stakes, fits in a weekend when the truck is between phases.
03Parked

Quiet Hours.

A small app that enforces real focus blocks across all my devices — no whitelist exceptions, no “five more minutes.” Built it for myself. Wondered if others would want it.

Why parkedApple and Google both ship adjacent features yearly. Building on someone else’s roadmap is a bad bet.
04Killed

Newsletter tool.

Briefly considered building a markdown-first email tool for solo operators. Started a prototype on a Tuesday.

Why killedKilled it Wednesday. The world has 40 of these. None of them are the bottleneck on anything I care about.
04 · The bench itself

Tools on hand.

Updated Apr 2026Boring on purpose.
Editor & daily
Editor
Cursor
Terminal
Ghostty
Notes
Obsidian
Tasks
Linear · solo
Browser
Arc still
Stack defaults
Frontend
React · Tailwind
Backend
Node · Postgres
Infra
Vercel · Neon
Auth
better-auth
Models
Whatever Patrick says
Hardware
Machine
M3 Max · 64GB
Display
Studio Display
Keys
HHKB Pro
Audio
Sonos · always on
Backup
Time Machine + offsite
Where the work happens
Primary
Home office
Mornings
5–8am · best hours
Afternoons
Calls + ops
Walks
For hard problems
Garage
For everything else
Note on the stackNone of this is precious. The stack is whatever lets me ship without thinking about the stack. If something better comes along that’s obviously better — not just newer — it gets swapped. If it’s just newer, it doesn’t.

Got something broken?

The bench has room. If you’ve got a problem that fits the work — and you’d rather build it than buy it — let’s talk.