The Vertical Writing Problem Is a Cross-Border Problem
Most cross-border operators I talk to think of localization as a translation layer: swap the copy, convert the currency, ship it. Then they hit Japan and discover that the medium itself is different. Japanese fiction isn’t just English fiction in another language — it’s vertical columns read right to left, ruby text (furigana) floating above kanji, manuscript paper conventions, and export formats that publishing workflows actually accept. Every general-purpose writing tool treats these as edge cases. That’s the gap Illusions is attacking, and while it looks like a niche writer’s app, the underlying lesson — that “edge case” features are often the entire market — is one every DTC brand expanding into Japan should internalize. This isn’t a review of a writing editor. It’s a case study in what happens when you build for the user everyone else deprioritized.
What Illusions Actually Solves (And Why It’s Not a Notion Clone)
The maker, Junpei Ikuta, frames the problem plainly: Japanese manuscript writing needs vertical writing, ruby, Japanese proofreading, and reliable manuscript export — and general-purpose editors treat all four as afterthoughts. Illusions bundles those with autosave, version history, and exports to PDF, EPUB, and DOCX, running across iOS, macOS, Windows, and Linux. That’s the whole pitch, and it’s a deliberately narrow one.
Compare that to the incumbents a Western operator would instinctively reach for. Notion is a database with a text box bolted on — beautiful for wikis, hostile to vertical typesetting. Google Docs can technically render Japanese, but ruby text and vertical layout are not native concepts; you’re fighting the canvas. Scrivener gets closer on manuscript structure but has never seriously served the Japanese fiction market. And the actual Japanese incumbents — Ichitaro, Word with vertical text plugins, dedicated manuscript software — are either aging, Windows-bound, or clunky in exactly the way Ando describes in the comments: “the old days where trying to write Japanese on a western PC was an absolute nightmare of encoding issues and clunky software.”
Illusions’ bet is that cross-platform consistency plus native vertical writing is a wedge no incumbent owns. That’s a real gap. Whether it’s a business is a different question, and I’ll get to that.
Why this matters more to marketplace sellers than to SaaS founders
Here’s the part most cross-border operators will skip past, and shouldn’t. The Illusions launch is a live demonstration of a principle that shows up everywhere in Japan market entry: the feature you consider a nice-to-have is frequently the reason the customer buys at all. Amazon Japan listings that ignore Japanese address formatting, sizing conventions, or seasonal gifting calendars underperform not because the product is wrong but because the presentation medium was never localized. The same logic applies to packaging inserts, instruction manuals, and customer service macros. Vertical writing is Illusions’ ruby text. What’s yours?
How It Differs From the Tools You Already Stack
Let me be precise about the comparison set, because “writing app” is a useless category.
- Notion / Coda: collaboration-first, block-based, zero vertical text support. Irrelevant to fiction manuscripts.
- Google Docs: the default, and the default is the problem. Ruby and vertical layout are not first-class.
- Scrivener: strong manuscript organization, weak Japanese typography, desktop-centric.
- Legacy Japanese manuscript tools: native features, poor cross-platform story, dated UX.
Illusions sits in the gap: native Japanese typography plus modern cross-platform delivery. The maker is explicit that GPT-6 Astra was used behind the scenes to review and refine the codebase, simplify architecture, and improve code quality — not to define the product. That distinction matters. The AI here is infrastructure, not the pitch. In a launch cycle where every other Product Hunt entry leads with “AI-powered,” Illusions leads with vertical writing and lets the AI be plumbing. I find that refreshing and, frankly, more honest about where AI actually creates leverage for small teams.
The tooling-stack parallel for operators
If you run a Shopify storefront expanding into Japan, your stack probably looks like: translation app, currency converter, maybe a localized Klaviyo flow. None of those touch the medium. The equivalent of Illusions in your world would be a theme layer that natively handles Japanese typography, address input, and payment method expectations (Konbini, PayPay) rather than a plugin that approximates them. That product mostly doesn’t exist as a turnkey, which is why Japan expansion is expensive. Watch for it.
What Cross-Border Sellers Can Borrow From This Launch
Three transferable plays, in order of how quickly you can act on them.
1. Build for the “edge case” that is actually the market. Illusions didn’t try to beat Notion at general note-taking. It picked a segment — Japanese fiction writers — where the incumbents are structurally weak. Your version: pick a market where the mainstream tooling is visibly bad (Japan, Korea, parts of MENA, Brazil) and localize deeply rather than broadly. A single well-served market beats five badly-served ones.
2. Let AI be infrastructure, not the headline. The maker used Astra for codebase review and architecture cleanup. That’s the highest-ROI use of AI for a small operator too: not customer-facing gimmicks, but internal quality and maintenance. If you’re running a lean Amazon FBA brand, point your AI budget at listing optimization, review analysis, and inventory forecasting — the unglamorous plumbing.
3. Treat export formats as a product feature. Illusions exports to PDF, EPUB, and DOCX because that’s what publishing workflows demand. For e-commerce, the analog is your data out: can your customers and partners get your catalog in the format they need? CSV, API, feed files for Google Shopping, TikTok Shop bulk uploads. Export flexibility is underrated retention.
Where the math breaks
Now the skeptical part. Illusions’ addressable market is Japanese-language fiction writers who want a cross-platform editor with native typography. That’s a real but narrow segment. Japanese fiction writing skews toward established desktop workflows and, increasingly, toward whatever the publisher’s submission system requires. A new editor has to win on workflow lock-in — version history, autosave, export fidelity — not just feature parity. The cross-platform promise (iOS, macOS, Windows, Linux) is expensive to maintain for a small team, and the maker hasn’t disclosed pricing or a monetization model in the launch material. That’s a yellow flag for anyone evaluating it as a long-term tool. If you’re a writer considering it, check the pricing and data-export story before you commit a manuscript.
For operators, the lesson cuts the other way: narrow segments are great for product-market fit and terrible for scale. Illusions may be a wonderful product that never becomes a big business. That’s fine for its users and irrelevant to yours. Don’t confuse “solves a real problem elegantly” with “a market you should enter.”
Where My Judgment Says It Falls Short
I’ll be direct. Illusions is a thoughtful product with a clear thesis, and I have three reservations.
First, the moat is thin. Vertical writing, ruby, and Japanese proofreading are features, not platforms. A well-funded incumbent could ship them. The defense has to be craft and community, and craft is hard to sustain cross-platform.
Second, the launch material is thin on business model. No pricing, no team size, no roadmap beyond the feature list. The comment from Dhiraj Wohra — “market yourself on Note, you should get more audience” — is well-meaning but points at the real gap: distribution. A great editor with no distribution is a hobby.
Third, the AI framing is a double-edged sword. Using Astra to clean the codebase is smart. But in a launch cycle dominated by AI hype, leading with “we used AI for architecture” risks reading as a footnote rather than a differentiator. The maker was honest about it, which I respect, but honesty about a non-differentiator doesn’t create one.
What I’d Watch / Test Next
If you’re a cross-border operator, the actionable move this week isn’t to download Illusions. It’s to run the audit Illusions represents. Take your top three non-English markets and ask, feature by feature, where your storefront, packaging, and support treat the local medium as an edge case. Japan is the obvious one, but Korea, Taiwan, and the Gulf have their own “vertical writing” equivalents — input methods, honorific conventions, payment rails. Then pick the single highest-friction item and fix it natively rather than with a plugin. If you’re a writer or content operator, watch Illusions’ pricing page and export fidelity before committing; the cross-platform promise is only as good as its maintenance. And if you’re building anything for a niche market, steal the playbook: pick the segment the incumbents structurally can’t serve, let AI handle the plumbing, and make the “edge case” the headline.
