AI adoption, web apps, booking and admin systems, and changes to services you already run. I work out what the real problem is, then handle design, implementation and production operation end to end. Bring a finished spec or a rough idea — both are welcome.
Start with whichever is closest. If none of them quite fit, describe the situation anyway and we will work out the fastest route together.
Built alreadyMUSUBU, the automated notifications in SHARENAIL, the rule-based input support in SUMI
AI is not a precondition. If plain automation is enough, that is what I will propose.
Built alreadySHARENAIL, Hanami Nail
This covers reviewing the implementation and the configuration. It is not penetration testing, not an exhaustive vulnerability assessment, and not a guarantee that nothing will go wrong.
Built alreadyMUSUBU, SHARENAIL
Built alreadyHanami Nail, SHARENAIL, this site (shingokumon.com)
I cannot promise that a generative AI will cite you, or that you will rank. What I can do is make the content straightforward to read and to quote.
Contract terms, start date and monthly rates are published on /en/work.
Explore their implementation details and public URLs.
Chair-rental matching platform for nail artists
Registration, screening, viewings, contracts and onboarding pulled into a single path, with the operator deliberately kept in the loop. Includes an admin dashboard and automated email triggered by the confirmed date (idempotent, dispatched every morning).
Related: Web apps & SaaS / Booking, LINE & matching
Beauty discovery platform for visitors to Japan
An English-first content platform built from scratch, structured so categories can go live one at a time. Implemented Sumi, a rule-based guide that suggests salons based on preferences and styles.
Related: Websites & findability / Web apps & SaaS
LINE × AI booking SaaS for beauty salons
Implemented AI handling of LINE enquiries, availability checks, reservation proposals and booking confirmation after customer approval.
Related: AI adoption & process automation / Booking, LINE & matching
SUMI Smart Inquiry — the enquiry form on this site — was built the same way. It reveals optional follow-up fields based on the keywords you type: rule-based input support, not a generative model interpreting free text.
Landing pages and small sites run ¥300,000 – ¥600,000 (2–4 weeks); business systems with an admin dashboard run ¥600,000 – ¥1,500,000 (1–3 months); SaaS and platforms: from ¥1,500,000 (3+ months). Maintenance: from ¥30,000/mo. Final scope moves the number either way. If you have a fixed budget, tell me — I'll propose what fits inside it.
Updates, backups, security, incident response, and improvement reports. I do not build and disappear.
Contract engineering — joining your team — has its own terms and monthly rates, published on /en/work. They are not duplicated here.
MUSUBU is a LINE × AI booking SaaS for the Japan market. This is product pricing, separate from the SMALL / STANDARD / PLATFORM tiers above.
Not a custom build — a product that is already running. AI is built into the LINE Official Account, automating everything from the first inquiry to a confirmed booking.
No per-booking commission. Built for salons paying heavy fees to booking portals.
The first message can be plain text. A call is not a requirement, and you will not be handed to someone else halfway through.
Describe what is getting in your way. A call is not required. I will ask about anything that is missing.
What gets built and what does not, stated separately. Scope, timeline and cost come before the work.
You see something running early and often. The person you talked to is the person who builds it.
Verified in the production environment, then released, with the operating notes you need.
Incident response and iteration after launch — on a monthly plan or job by job.
A. Not at all. We start from where the time goes and where things fall through the cracks. Most of the conversations I have begin at that stage.
A. Yes. I start by reading the code, isolating the cause and agreeing what is in scope. Reviewing auth, permissions and RLS, and updating dependencies, can be part of it.
A. No. If automating notifications and reporting is enough, that is what I will propose. An LLM or RAG setup comes in when you need search and answers grounded in your own data.
A. I review the implementation and the configuration: auth and permission design, Row Level Security, how input is handled, and dependency updates. I do not run penetration tests or exhaustive vulnerability assessments, and I do not guarantee that nothing will go wrong.
A. Everything is done online. I do not work on-site. Calls happen during Japan business hours; the rest of the work is async.
Write down what is taking up your time right now. We can work out what is worth building from there.
Send any time — I reply within 1 business day.
Contract rates and availability are published on /en/work.
By email: info@shingokumon.com