I built a self-growing Japanese brand database in 1 hour with Lovable (0 customers, 0 followers)

#ai #buildinpublic #lovable #showdev

TL;DR: Japanese Amazon seller from Osaka. Yesterday I built Japan Brand Finder in 1 hour with Lovable — a tool helping English-speaking Amazon sellers source from Japan. Currently 0 customers, 0 followers. Building in public. Demo

Why I built this

I'm based in Osaka and I run an Amazon Japan business. For years I've watched English-speaking sellers struggle to source from Japan — language barriers, FAX-only wholesalers, and the silent fear of writing emails that accidentally sound rude.

So I sat down with Lovable and built the thing I wish existed when I started.

What it looks like

The pitch on the landing page: "Search a growing database of Japanese makers — from Tsubame-Sanjo metalworkers to Kyoto matcha houses. Get wholesale contacts, MOQs, English-support level, and culture notes. Then let AI write the polite outreach email for you."

I didn't write that copy. Lovable did. I never mentioned Tsubame-Sanjo (a famous Japanese metalworking region) in my prompt — yet the AI surfaced exactly the kind of specific, credibility-building example that resonates with Amazon sellers who source from Japan.

The database that grows itself

Here's the thing I'm most proud of.

Look at this screenshot. Kyocera, Zojirushi, Tiger Corporation, Daikin, Sharp, Suntory, Nissin Foods, House Foods, IRIS Ohyama. 9 Japanese brands. All marked "AI" — meaning I didn't seed any of them.

When a user searches and the brand isn't in the DB, the AI generates a real entry — name in English and Japanese, category, HQ location, English support level — and saves it to the shared database. The next user who searches that brand gets an instant hit.

Day 1: 20 seed brands
Day 2: 30+ brands and growing
Every user makes the product more valuable for everyone

This is the moat. Not the UI. Not the features. The DB.

How good is the AI-generated outreach email?

This is one click away. Hit "Generate" → AI writes a polite Japanese business email like:

件名：御社製品の卸販売に関するお問い合わせ（海外Amazon販売店より）

池永鉄工株式会社御担当者様

突然のご連絡失礼いたします。

海外にてAmazonストアを運営しております、久保川護と申します...

Three things to notice:

Subject line auto-generated — most AI tools forget this
Native keigo — "突然のご連絡失礼いたします" is the textbook polite opener
Real product knowledge — the email mentions "鉄なべ" (iron pot), Ikenaga's actual product line. The AI pulled this from the brand's business culture metadata

For an English-speaking Amazon seller trying to source from Japan, this single feature replaces a $200/hour translator.

The 1-hour journey

Time	What happened
0:00	Wrote a 200-word spec
0:01	Pasted into Lovable
0:08	First MVP shipped (search, details, favorites, AI email)
0:30	Multi-language search bug fix prompt
0:55	3-mode email feature added with before/after diff
1:00	Free credits ran out at exactly the right moment

What I learned

1. The "Both" option is gold. When Lovable asked how to populate the brand database, the answer was "seed sample brands + AI enrichment on misses." This single decision is what turns the tool into a growing asset.

2. Security scan saved me. Before publishing, Lovable's automated security scan flagged a critical issue (RLS missing — wholesale contacts would have been publicly readable). One click fixed it. Without that scan, I would have shipped a privacy violation.

3. Honest copywriting from AI surprised me. The Tsubame-Sanjo example. The 鉄なべ reference. These weren't in my prompts. The AI just knew what would resonate.