Skip to content

Systems we’ve already built, and what they prove.

Before we build for clients, we build for ourselves. First, a showroom of three flagship websites you can open and explore. Then the products we designed and shipped end to end: the problem each one solves, what we built, where AI does the work, and the stack underneath. The same approach goes into the systems we build for you.

Flagship Website Showroom

Live demoOpen the showroom ↗

Three flagship websites for three very different businesses, shown on 3D laptops and phones you can rearrange, open, and step inside.

The problem

A screenshot cannot show what a flagship website feels like, and “trust us” is not proof. You should be able to open the work and use it before the first conversation.

What we built

  • A 3D showroom that runs in the browser: three laptops, each with a phone beside it, under its own spotlight
  • Laptops you can drag into a new order, with lids you can pull open or closed
  • Zoom all the way into a screen, or click a screen or its phone, to step inside that site
  • Solmère House, a luxury hospitality site: fourteen suites above a private cove, told as one day from first light to night
  • Relay Nine, a same-day courier site: an instant quote form, one delivery followed from order to signature, and a live board
  • Halden & Pike, a plumbing service site: the emergency line at the top of the page, upfront pricing, and a direct path to book
  • Each concept site is labeled as a concept, and “Book a call” leads to our discovery form

What to try

Drag the laptops into a new order. Pull a lid shut and open it again. Then zoom into a screen, or click it, and you are inside the site itself, with its own look, voice, and path to a booking. The three businesses are fictional. The sites are built the way we would build one for you.

Built with

  • Three.js
  • WebGL

What it proves: A flagship site does not have to look like ours. Hospitality, logistics, and a local trade each get their own design, built to the same standard.

ResellerScore

Liveresellerscore.com ↗

Photograph a thrift find, get an AI resale value backed by recent sold comparables, then turn it into marketplace-ready listings from one inventory record.

ResellerScore inventory screen showing revenue, profit, return and items sold, with profit by category and tracked items.
ResellerScore home screen with tracked inventory, open listings, sold items and the recommended next step.

The problem

Resellers guess at value on the shelf, rewrite the same listing for every marketplace, and risk selling one item twice.

What we built

  • Multi-photo AI identification and condition-aware valuation, plus barcode scanning
  • Live web research on recent sold comparables, with sell-through and days-to-sell shown only when the evidence supports them
  • A buy-or-pass sourcing decision and profit scorecard
  • AI-written listings optimized for each marketplace: its own title, description, price, and structured fields, all editable
  • One inventory record across 11 marketplaces: a direct eBay integration (beta), a browser helper that fills in Facebook Marketplace, Poshmark, Mercari, and Etsy listings, and copy-ready listing kits for six more
  • Recording a sale lowers the quantity and queues delisting everywhere else in one step

Where AI does the work

Vision AI identifies the item and grades condition; an AI research step pulls recent sold comparables from the web and returns structured, evidence-backed pricing; then AI writes a listing optimized for each marketplace: the title, description, price, and fields that marketplace rewards. A second AI provider stands by as a fallback, and an image AI cleans up listing photos.

Built with

  • OpenAI
  • Anthropic Claude
  • Photoroom
  • Next.js
  • Clerk
  • Stripe
  • Postgres
  • Vercel
  • Chrome extension

What it proves: AI that turns messy real-world input (photos) into structured records your team can act on.

LocalDeliveryOS

Early accesslocaldeliveryos.com ↗

The operating system for regional delivery companies: order entry, dispatch, proof of delivery, billing, payments, and driver pay in one connected record.

LocalDeliveryOS illustrative order journey: one delivery followed from order entry through proof of delivery, invoice, payment, driver pay and margin.
LocalDeliveryOS flagging a completed delivery with no matching invoice line as a potential unbilled delivery for a person to review.

The problem

Completed deliveries fall through the gaps between dispatch, billing, and driver pay, and work that never gets invoiced is revenue that quietly disappears.

What we built

  • Order intake, both manual and through a customer portal, plus a live dispatch board
  • A driver phone app with signature and photo proof of delivery, and a location check that flags arrivals logged far from the address
  • Public tracking pages for customers
  • Configurable pricing: accessorials, waiting time, after-hours, fuel, and customer-specific rates
  • Delivery-linked invoicing on each customer’s billing cycle, and driver settlements across six pay types
  • A revenue-leakage audit that flags completed deliveries with no invoice line, for human review

Where AI does the work

Deliberately rules, not AI, where money is decided: billing exceptions are caught by deterministic checks, each with a plain-English “why was this flagged?” explanation. We use AI where judgment helps, not where a rule is more reliable.

Built with

  • Next.js
  • TypeScript
  • Supabase
  • Stripe Connect
  • Google Maps
  • QuickBooks Online
  • Resend
  • Vercel

What it proves: One connected record from first order to final payment, so finished work always becomes revenue.

NEMT Logic

Betanemtlogic.com ↗

The financial operating system for non-emergency medical transport (NEMT) providers: dispatch, drivers, billing, and settlements in one workspace.

NEMT Logic operations view: reassign trips, a live map of active drivers, driver day view and equipment-aware assignment.
NEMT Logic illustrative revenue and margin panel with billed, collected and margin by payer.

The problem

NEMT companies run on phone calls and paper, and they live or die in accounts receivable, not in dispatch.

What we built

  • A dispatch board that matches trips to vehicle equipment (wheelchair, stretcher, or ambulatory) with recurring and will-call trips
  • A browser-based driver app with electronic proof of delivery and vehicle inspections
  • An append-only trip history, so every change is traceable
  • Instant trip quoting from versioned payer rule packs, and a 10-rule pre-claim checker
  • A facility and broker portal with live tracking links
  • Invoices with card payment, and driver pay for both W-2 and 1099 drivers

Where AI does the work

Automation, not AI: payer rules and pre-claim checks run as versioned, testable rules, because billing accuracy has to be provable.

Built with

  • TypeScript
  • Node.js
  • Postgres
  • React
  • Clerk
  • Stripe Connect
  • Resend
  • Railway
  • Vercel

What it proves: Complex, rule-heavy billing turned into a system a small team can actually run.

Staff HQ

Private betastaffhqos.com ↗

AI staff, organized: people and AI specialists working together in group chats to get real work done, with a person assigned ultimate responsibility for every task and project.

Staff HQ command center showing overdue and upcoming tasks for the day, with AI staff roles listed in the sidebar.

The problem

AI help is scattered across a dozen private chat windows, cut off from the rest of the team, with no shared context, no visibility, and no one clearly accountable for the result.

What we built

  • Group chats where people and AI specialists collaborate on real work, sharing the same context
  • Assign a person ultimate responsibility for any task or project: AI assists, a human owns the outcome
  • A Chief of Staff agent that asks clarifying questions, then coordinates seventeen specialist roles you can switch on or off
  • Projects with saved knowledge, so context carries over
  • Gmail review with drafted replies, plus calendar help; nothing sends without your explicit approval
  • Voice dictation and scheduled jobs
  • Activity, usage, and spending controls

Where AI does the work

AI specialists work alongside people in the same group chats, drafting, researching, and following up, while a named person owns each task and project. Routine and sensitive work are routed to different AI models, requests are sent with provider-side storage turned off, and a person approves every outbound action.

Built with

  • OpenAI
  • Next.js
  • Clerk
  • Supabase (row-level security)
  • Gmail & Calendar APIs
  • Vercel

What it proves: People and AI working as one team, with a human accountable for every outcome: the model we bring to client teams.

AI Margin

LiveWatch on YouTube ↗

A YouTube news channel run entirely by AI. It posts a scheduled video every night and short videos through the day as AI news breaks, with no person in the loop.

The problem

AI news moves faster than a person can script, record, edit, and post. By the time a traditional video is finished, the story is old.

What we built

  • Continuous monitoring of the web for new AI stories
  • A filter that picks out the stories with real impact on businesses
  • Script writing, start to finish
  • An AI-generated host who presents each story
  • Final video processing
  • A scheduled nightly video, plus multiple shorts through the day as news breaks
  • Automatic posting to YouTube

Where AI does the work

Every step. AI watches the news, decides what is worth covering, writes the script, generates the host, assembles the video, and posts it. Shorts go into production as soon as a story breaks, the nightly video runs on a schedule, and no person touches either before it is published.

What it proves: A complete process, from trigger to finished product, running with no manual steps. On client work that touches customers or money we add a human approval point. This shows the pipeline underneath can run on its own.

Private AI Second Brain

Private, in daily use

A private AI command center we built to run our own day (calendar, inboxes, voice notes, tasks, and projects) from one dashboard and assistant.

The problem

Context scattered across mailboxes, calendars, a voice recorder, and many projects, with no time to pull it together.

What we built

  • A “start my day” briefing
  • Inbox triage into Needs reply, Waiting, and FYI, with AI-drafted replies that only send after explicit confirmation
  • Voice notes captured into a journal, with commitments turned into reminders
  • A multi-calendar view where the assistant proposes events instead of booking them on its own
  • A project board with AI-suggested next steps that cite their evidence
  • Secure remote access from a phone, behind multi-factor authentication

Where AI does the work

An AI assistant uses tools on your behalf, and anything risky requires confirmation. Dictation is transcribed on the device itself, so only text ever leaves the machine.

Built with

  • Anthropic Claude
  • Python
  • SQLite
  • Gmail & Calendar APIs
  • On-device speech-to-text
  • Cloudflare Access

What it proves: A true second brain: the same kind of AI assistant we build for owners who are buried in email and scheduling.

What this looks like in practice

These are examples of the kinds of systems we build, not a list of finished projects, every one is shaped around the business it’s built for.

  • A custom company operating system

    Runs the core of the business (jobs, scheduling, approvals, and reporting) from one connected place instead of a dozen disconnected tools.

  • Customer and partner portals

    Gives customers and partners direct, branded access to status, documents, and updates, so they stop calling to ask what you already know.

  • Workflow and approval applications

    Replaces text-message and email approvals with a clear, trackable process that shows exactly where every request stands.

  • Field-service or delivery systems

    Connects dispatch, field crews, and customers in real time, so jobs are tracked from assignment to completion without a phone call in between.

  • Revenue-leakage detection systems

    Flags completed work that hasn't been billed and pricing that's drifted from policy, so revenue stops disappearing quietly.

  • Inventory and asset workflows

    Tracks materials, equipment, and assets as they move through the business, cutting down on shortages, overordering, and manual counts.

  • Scheduling and dispatch tools

    Matches jobs to the right people and equipment automatically, reducing the daily scramble to figure out who's doing what.

  • Internal knowledge and training systems

    Turns what currently lives in a few employees' heads into a system anyone on the team can reference and follow.

  • Custom dashboards and decision-support tools

    Puts the numbers owners actually need (margin, capacity, bottlenecks) in front of them without a manual report-building process.

Want a system like these for your business?

Start with a free Business Systems Mapping Session. We’ll map how work moves today and show you what we’d build first.