July 2026 — Earthquake risk built in, a one-step trial, and the market map goes GA

April was about data, May about personalization, June about judgment. July was about the two questions every investor asks before anything else: how risky is this, really? and how fast can I see for myself? We built earthquake risk into every property using Japan's official seismic hazard model, took the Market Intelligence map out of beta with drill-down to the municipality, and rebuilt the first five minutes of the product — sign up, and a fully analyzed property is already waiting for you. Here's what shipped.

1. Earthquake risk, from Japan's official seismic hazard model

Every property's hazard assessment now includes earthquake — the sixth hazard type, alongside flood, tsunami, and the rest. We draw it from J-SHIS, the national seismic hazard model published by NIED, Japan's disaster-science research institute: for each property you see a clear hazard tier and the headline figure behind it — the probability of strong shaking (JMA intensity 6-lower or higher) within the next 30 years.

We were careful with the framing, because raw earthquake numbers mislead. Much of urban Japan — including most Pacific-coast metros — carries a high baseline probability, and that's normal; what separates a sound purchase from a risky one is not the hazard figure alone but how the building answers it. So the rating is presented the way a professional would present it: hazard, then mitigation — and the AI analysis now reasons about the two together, pairing the local probability with the building's seismic-standard era (新耐震 / 旧耐震) to tell you what the number actually means for this building. There's also a new seismic hazard overlay on the property map, so you can see how shaking probability varies across an area at a glance.

Source: J-SHIS seismic hazard model — NIED (National Research Institute for Earth Science and Disaster Resilience)

2. The Market Intelligence map leaves beta — and drills down to the municipality

The prefecture market map introduced in May is now generally available for every user, and it goes a level deeper: click a prefecture and it opens into a municipality-level choropleth, with a detail drawer for the city or ward you select. Properties are now matched to municipalities precisely — using official JIS municipality codes rather than address-text guessing — so filtering from the map to your properties lands on exactly the right ones.

The same municipal data now feeds the AI: property analyses cite market context for the property's own municipality — transaction activity and price signals for the actual city or ward — instead of stopping at the prefecture.

Source: MLIT transaction and land-price data + IPSS demographic projections, keyed to official JIS municipality codes

3. Sign up, and a property is already waiting for you

Starting a free trial used to walk you through a checkout page. That step is gone: choose your plan and you're in — no checkout, no card. And the first thing you see is no longer an empty workspace: every new account comes with a demo property and a complete showcase analysis, so you can explore what a full Tatemono IQ analysis looks like before you've imported anything of your own. A new getting-started checklist then walks you through the steps that matter — importing your first property, running your first analysis — and property setup got lighter touches too, like auto-generated property codes.

4. A new analysis type: Ryokan (旅館業法)

Some properties aren't 180-day minpaku candidates — they're candidates for a full hotel/inn license (旅館業法) and year-round operation. There's now a dedicated Ryokan analysis type for exactly that. It models the 365-day operating economics (room rates × occupancy, staffing and service costs), the licensing path — a permit from the public-health office, not a minpaku notification — zoning eligibility, and the one-time conversion investment a facility-standards and fire-code upgrade typically requires, then puts the result side by side against minpaku and long-term rental so you can see which strategy the property actually supports.

5. Maisoku export, redesigned

The Maisoku (マイソク) PDF export got a professionally redesigned default template: fields prioritized the way standard Japanese property sheets order them, a compact location map up top, and broker/team logos rendered at their correct proportions. Address privacy is now enforced end to end — when a property's exact address is restricted, every part of the export respects it: the map shows an area boundary instead of a pin, and the printed address and AI-written synopsis stay at area level too.

6. Import from Hatomark Site, and a more forgiving extension

The browser extension now supports hatomarksite.com (ハトマークサイト), the listing portal of Japan's national real-estate association network — bringing one-click import to another major listing source. Imports also got more forgiving: a listing with an odd or incomplete field now saves as a draft for you to finish in the app instead of failing outright, and floor counts are read correctly on buildings with basement levels. The Chrome Web Store listing was polished too, with proper screenshots and localized descriptions in all five languages.

7. Help, right where you need it

The product documentation now lives inside the app — no more switching to the website to look something up. We added new guides for Teams & Roles and Inviting Investors & Sharing Properties, and Aki, your AI assistant, learned the documentation too: ask which listing sites the extension supports, or where to find a guide, and Aki answers and points you to the right page.

Also improved in July

• The price you see is the price you're billed. Signup and checkout now carry your language and displayed currency end to end — including on resumed subscriptions, which previously could re-bill in a stale currency.
• Smoother sign-in: email addresses are no longer case-sensitive at login, sign-in errors are actually shown instead of failing silently, and email-verification screens now report exactly what happened.
• AI analysis polish: the narrative and the numbers tables can no longer disagree on figures like net yield, internal codes and model self-talk no longer leak into the output, and cost-approach (積算) valuations are computed correctly for sectional-ownership units.
• The AI engine was upgraded — property analyses and Aki now run on a newer generation of Google's Gemini models, with faster responses and sharper reasoning.
• The property form no longer loses unsaved input when you switch languages.