Lonely Pine AI Collaborator Brief
What we build, and where you'd start

One studio. Eight production systems. Built with AI, end to end.

A walkthrough of Lonely Pine AI: the live client systems, the architecture underneath them, and the three projects you could start on.

Prepared for
Amit
Prepared by
Nalu, Lonely Pine AI
Date
August 2026
First project
To scope together
01 / What Lonely Pine is

AI handles transactions so humans can handle interactions.

Lonely Pine AI is a founder-led studio based in Tokyo, serving clients in Hawaii and beyond. After a decade at a Fortune 10 company, Nalu left in June 2025 to build AI systems for real operators: restaurants, vacation rental managers, boutique resorts, painting contractors. Not demos, not decks. Production systems that answer phones, write orders, and run back offices every day.

4
Service lines
Voice agents, workflow automation, intelligence dashboards, and fractional Chief AI Officer engagements.
5
Live voice agents
Five AI agents on real phone numbers today: food ordering, guest services, lead capture, outbound sales, reception.
1 + AI
The whole team
One founder, AI-augmented from design to deploy. That is the operating thesis, and it works.
02 / Shipped and live

Real clients, real production.

Every system below is deployed and in use today. The architecture line under each card is the part you asked about. If a term on it is new to you, that is fine: section 03 walks the whole stack from the top.

Restaurant · 3 locations

Sparky's Food Co.

Hoku, an AI phone agent that takes food orders on the main line, prices them from the live Square catalog, writes real orders into Square, and texts payment links. Around it: an order tracker that reconciles phone revenue against real payments, a review detection and response layer, and a Telegram control the owner uses to pull an item off the menu, which moves the database and the live voice prompt in one tap.

StackVapi voice · Cloudflare Workers (TypeScript) · Square API · D1 · Telegram bot · Twilio SMS
Painting contractor · Fractional CAO

Equipped Painting

Steward, a Claude-powered operations platform the whole company uses: pipeline intelligence rebuilt on a reconciled data warehouse, a knowledge graph of the org where document sensitivity inherits role visibility, 21,073 jobsite photos mirrored and joined to the production board, weekly scorecards into Slack, and Quinn, their outbound sales voice agent. A daily job asks the assistant ten questions and checks each answer against the database, so a wrong number pages us instead of reaching an owner.

StackClaude API · Retell voice · Workers + D1 + Vectorize · DripJobs, CallRail, CompanyCam · Slack · 788-test suite
Beach resort · Voice + ops

Kamaole Beach Club

Malia, a guest-facing voice agent that handles arrival questions and weekly check-in logistics, plus an operations dashboard behind magic-link auth. Underneath it, a full interval-ownership model: 1,617 deeds, 839 owner entities reconstructed from a source that carries no owner key, and a live review queue where staff resolve what the data cannot decide on its own.

StackVapi voice · dedicated Worker + D1 · magic-link auth · deed registry + review queue · Pages frontend
Vacation rentals · 72 units

DMV Maui

Lani, a voice agent that quotes live availability sourced from the real booking site, refreshed by a pre-warming cron across all 72 units. Plus a daily operations board for the front desk where the build fails if a door code, pool code or Wi-Fi password ever appears in the output, a social publishing queue, automated newsletters, and a monthly reporting suite.

StackVapi voice · Workers + D1 + KV + R2 · Cloudflare Access · build-time redaction tests · Mailchimp API
Hospitality group · Document AI

DMI / Blu

Blu, a document-intelligence assistant for a hospitality group managing about 50 associations: ingestion pipelines over their file corpus, threaded conversations, and locked-template reports on scheduled crons. Its document editor returns a list of edits, each carrying a fingerprint of the text it expects to find, and if one anchor fails to match, nothing is written at all.

StackClaude API · Workers ingestion · Pinecone + R2 · anchored DOCX engine · cron reporting · email distro
Resort condo · Ops dashboard

Maui Eldorado

Operations dashboard on a real database: guest check-in forms routed straight into D1, client-editable tour content, competitor-listing suppression on a scheduled job, and an automated monthly owner report. Guest broadcasts fan out one message per guest, which is how a tropical storm advisory reached the property when a single bulk email would have been rejected outright.

StackPages + D1 · form-to-database routing · per-guest batch email · external policy canary
Tokyo church · Pro bono · Mobile

Franciscan Chapel Center

A native mobile app for an international parish, live on the public App Store since July and in tester rollout on Android. Mass and confession schedules, ministry rosters, a weekly bulletin archive, announcements in six languages, a beta feedback pipeline writing into a database, and admin dashboards for the parish office.

StackReact Native + Expo · EAS builds · Workers API + KV + D1 · 6-language bundles · Pages admin
Internal · The cockpit

LP Command Center

The studio's own operating cockpit, on the desktop and now as a native iOS app on TestFlight: daily pulse across every client system, task boards, and an ingest pipeline that captures anything saved into a Slack channel, stores it first and classifies it second. The phone carries a scoped device token, never the master key.

StackExpo iOS app · Workers + D1 + R2 · Cloudflare Access · Slack events + Haiku classification

Also live: Gabby, Lonely Pine's own receptionist, answering questions about the business from a searchable corpus of its documents; hawaiicustomtreasures.com, a pro bono small-business site; and Wayfinder AI, the studio's own newsletter and funnel.

03 / Under the hood

One architecture, many clients.

Everything runs on the same layered pattern. A client engagement is mostly configuration and integration on top of proven plumbing, which is why one person can run eight production systems. Each layer below says what it is for first, then what it is built with.

Voice layer
The front door

What it does: answers and places real phone calls. The customer talks to an AI agent instead of a menu tree. Each agent has its own script, its own list of things it is allowed to do, and rules for how it says numbers out loud, because a phone number read as a figure is useless to the person writing it down.

Why it is a layer: the agent knows how to hold a conversation but nothing about the business, so mid-call it calls our own backend for live answers.

VapiRetellTwilio numbersTTS pronunciation rulessilence + idle handling
Agent intelligence
The brain

What it does: the actual reasoning. Claude runs the conversations, reads the documents, drafts the reports, and answers questions about a client's own files by searching them first and citing what it found.

Why it is a layer: the model is given tools rather than trusted from memory. It searches, it computes, and every fact on a client surface traces back to something we can point at.

Claude APItool callingVectorize embeddingsretrieval over client corporasensitivity tagging
Backend layer
The plumbing

What it does: the code that actually runs. Small TypeScript services deployed to servers we never manage, plus scheduled jobs that fire on their own: nightly reports, hourly syncs, health checks.

Why it is a layer: it started as one shared service and has been split into a dedicated one per client, so a bad deploy for one client cannot touch another. Anything automated fails loud. A broken report pages us; it never silently skips.

Cloudflare WorkersTypeScriptcron triggersper-client isolationautomated test runner
Data layer
The memory

What it does: remembers things between calls and between days. Three kinds: a real SQL database for records, a fast key-value store for settings and cached answers, and a vector store that finds documents by meaning rather than by exact wording.

Why it is a layer: every client gets their own database with ownership checks, so one client's data has no path to another's screen.

D1 (SQLite)KVVectorizeR2 object storageper-client databases
Integrations
The nervous system

What it does: talks to the software the client already pays for, in both directions. Their point of sale, their CRM, their call tracking, their email, their chat.

Why it is a layer: this is where most of the real difficulty lives. Vendor systems rewrite their own history, truncate their exports, and report failure as success, so the interesting work is deciding what to trust and proving it.

SquareDripJobsCallRailCompanyCamMailchimpSlackGoogle APIsStripeTwilio
Frontends + auth
The surfaces

What it does: the screens people actually use. Staff dashboards, owner consoles, guest forms, and native mobile apps when the product calls for it.

Why it is a layer: every screen is gated. Some by company sign-in, some by a one-time link emailed to a guest. Who can see which rows is a decision made in the backend, never in the page.

Cloudflare PagesCloudflare Accessmagic-link authReact Native + Expo
Build system
The multiplier

What it does: the entire studio is built and operated through Claude Code. Design, implementation, tests, deploys and reporting all run as deep AI-collaborative sessions, with quality rules enforced by machine rather than by memory.

Why it is a layer: this is the actual product of working here. It is the part most shops have not figured out yet, and it is the skill that transfers to anything you do next.

Claude Codeenforced brand + content rulessession continuity systemincident write-ups
Why this matters: none of these clients could justify a traditional dev team. The stack above keeps marginal cost near zero, so a solo studio can ship and operate systems that normally take a department.
04 / In flight right now

What's being built this month.

01Sparky's closure guard. The shop closed for three days and the voice agent kept selling food, because the point of sale can express a weekly schedule but not a one-off closure. Scoped and evidenced, not yet written.
02Equipped attribution, end to end. Connecting the assistant's marketing funnel to the reconciled warehouse so every lead traces to the channel that produced it, on a data source that quietly edits its own history.
03Command Center capture pipeline. Closing the last step so anything saved on a phone gets transcribed and enriched automatically instead of sitting as a link.
04Better retrieval on the company brain. Combining meaning-based and keyword search, then grading both against a fixed set of questions so an improvement can be proven rather than felt.
05Productizing the org brain. Pulling the shared spine out of two client builds into one library, so the next engagement starts at week six instead of week one.
05 / Where you'd start

Three real projects. We pick one together.

Rather than hand you a task, here are the three live tracks with room on them right now. They are genuinely different kinds of work, so the choice is really about what you want to get good at. We scope the first piece on the call.

Pro bono · Mobile

Franciscan Chapel Center

The Tokyo parish app, already live on the App Store with real users. The next build is a schedules and assignments layer for ministries that currently open to an empty screen. Self-contained mobile work with people waiting on the other end of it.

You'd touchReact Native + Expo · a Workers API · D1 · app store release flow
Client work · Design to launch

Halehaku Maui

A single-page site for a family art gallery and farm shop on the road to Hana. Small on purpose: it is the entire arc of a client engagement, design through launch, compressed. Nalu meets the family next week, so the start date is genuinely open.

You'd touchHTML + CSS from a design system · Cloudflare Pages · DNS cutover · on-page SEO
Client work · Data + ops

Kamaole Beach Club

Operations tooling for the resort's front office. The nearest piece is letting staff upload a two-page schedule instead of one, which sounds small and is actually the thing standing between the client and running this herself.

You'd touchTypeScript on Workers · D1 schema + migrations · file upload to R2 · a Pages dashboard
Your seat: you take as much as you want to take, starting small enough to finish. You will not be handed a ticket and left alone with it. We work the same session together, you see how a real engagement runs, and the piece you ship goes live with your name on it.
06 / How we'd work

We build it together, and it goes live.

01Pro bono to start, real credit always. Everything you ship is production work with your name on it. A portfolio of live systems, not tutorial projects.
02You do not need to already know this stack. You need to be able to read code and want to build. The studio's real skill is the AI-augmented workflow itself, and that is learned by doing it on something real, not by studying it first.
03Async and flexible. Scoped pieces you pick up on your own schedule, sized so each one is finishable. No standing meetings.
04A real path to paid work. As fit develops, paid project work is on the table. The clients above all started as one small thing that went well.

Let's put your name on something that's actually running.

Lonely Pine AI · JDKC Ventures LLC · nalu@lonelypine.ai