What Is UITruth? A Japan-Focused LLMO Tool for Measuring How You Appear in AI
Instead of searching "good labor-and-social-security office in Shibuya," people now ask ChatGPT to "recommend a good labor-and-social-security office in Shibuya." That's not a rare scene anymore.
When they do, does your name show up in the AI's answer? If it does, what's the context? How are your competitors handled?
Until now, there's been almost no way to check that — how you actually appear in AI. UITruth is built to measure exactly that.
What UITruth does
UITruth captures real, web-searching ChatGPT answers to Japanese-language questions and measures how your brand appears within them. Other AI systems are not supported today.
You can open ChatGPT yourself and read an answer in the moment. But recording that by hand every time, varying the question, comparing against competitors, and tracking how it changes over time — that's not realistic to do manually. UITruth turns that whole process into a system.
Here's what you can measure:
- Brand mentions: how often you and your competitors are referenced within the answer
- Cited sources: the websites and URLs the AI leaned on to build its answer
- Related questions (fan-out): the related searches the AI branched into from a single question
- Local results: the shops and offices that surface for a question with a place name
- Products: products that appeared in the answer, and their prices
- Ads: ads shown within the answer
Being able to track all of this at once, and continuously, is what sets this apart from just opening ChatGPT and glancing at it now and then.
Capturing "the real ChatGPT answer," not a summarization API
This part is worth explaining carefully.
UITruth does not read an answer back from a summarization API. It captures the body text and citation structure of a real, web-searching ChatGPT answer. When a measurement did not go through our default route, the result page says so.
UITruth sends the same question a real user would type to the real, web-searching ChatGPT. From the answer that comes back, we pull the body text plus the cited sources, related questions, and whether ads appeared. What we capture is the answer's text and structure — not a screenshot of the screen. If we're going to measure it, we wanted to measure it under conditions close to what a user actually sees.
LLMO, GEO, and AEO
It might be an unfamiliar term.
LLMO (Large Language Model Optimization) and GEO (Generative Engine Optimization) describe work to make a brand or its information easier for generative AI to use appropriately. AEO (Answer Engine Optimization) focuses on making information easy to find as an answer in search and AI answer surfaces. The terms overlap, but UITruth's role is measurement.
ChatGPT is the default AI UITruth measures (Gemini, real UI, is also available). It records web-searching, Japanese-language answers so you can compare what appeared before and after a change.
What changes, and for whom
Here's the kind of shift we mean:
- Before: search your own name in ChatGPT, get an impression in the moment, think "guess we're not showing up," and leave it there
- With UITruth: which questions you appear for and which you don't — compared against competitors — becomes a record you can act on
If you're in PR, you can check on a regular cadence how the AI is describing your company. If you're in marketing, you can compare share of voice against competitors. If you run e-commerce or a physical shop, you can track how your products and local listings appear.
It's not limited to companies and shops — you can measure a municipality, a region, an organization, or an individual or creator's name the same way, seeing how each is handled within AI answers.
Clear pricing, in yen, nothing added at checkout
UITruth prices everything in Japanese yen. No converting a dollar figure at month end, and nothing is added to the listed price at checkout.
- Free: weekly auto-tracking of 1 question for your first 14 days (it pauses after that; everything measured stays) + up to 15 checks a month. No credit card required
- Starter: ¥2,980/month · 5 weekly questions · up to 100 checks/month
- Pro: ¥9,800/month · 20 weekly questions · up to 300 checks/month. Parallel monitoring of multiple brands and competitors
- Agency: ¥29,800/month · 100 weekly questions · up to 1,000 checks/month
Start free, and pick a plan sized to your volume once you want to keep going. See the pricing page for details.
Just look at how you appear today
"How does ChatGPT actually describe you" is faster to just look at than to imagine. It's usually not what you expected.
That's the thinking. Here's the real thing
What UITruth captures is the real ChatGPT answer, not a summary of one. Your first run takes no account and no credit card. Decide whether it's a fit after you've seen it work.