2
1 Comment

Show IH: 70-year-old indie hacker from Japan, just shipped my first product (and second, in Japanese)

Hi Indie Hackers,

I'm smile3, 70 years old. I run a small company in Japan called SMILE70 LLC.
Yesterday I shipped my first two products on Gumroad — within 48 hours of each other.

Background

I'm building toward "Yumoa Village" — a network of 47 retirement communities across Japan over the next 10 years. To fund it, I run 4 AI businesses solo:

  1. AI Secretary (B2B for SMBs)
  2. Video Automation Studio (full-auto YouTube pipeline)
  3. IoT Property Monitor (vacant-house security)
  4. Education Store (this — selling what I learn)

All of these are powered by Claude Code (Anthropic's CLI agent).

What I shipped

Product 1: "Claude Code for Total Beginners — From a 70-Year-Old Solopreneur" ($29)

  • 130+ pages, 13 illustrations, scripts, templates, bonus skills
  • 5-layer copy protection (zero-width Unicode watermarks)
  • Beginner-friendly, every term explained, every command copy-paste

Product 2: Same, in Japanese ($29) — for Japanese expats and language learners

Why I made it

For 6 months I've watched young people install Claude Code, get overwhelmed, and quit.
Most "Claude Code beginner guides" are too technical, too fast, written in monospace fonts.

I wrote one for me — for the 70-year-old me who used to be intimidated by git push.
And I gave it the pace I'd want.

Numbers

  • Total time to write: ~3 weeks of "calm mornings"
  • Time from first draft to published Gumroad: 48 hours
  • Cost: $0 (Gumroad takes 10% on sale)
  • Target: $1k/month MRR by August 2026

What I learned

  1. The "70-year-old who runs AI businesses" angle resonates. People remember it.
  2. Bilingual product pages get more traffic (English on top, Japanese below — Gumroad's vector search picks up both).
  3. Don't underestimate copy protection — even just stating "this is a licensed copy" deters re-uploads.
  4. Existing audience matters. I had ~200 followers on note (Japanese platform) before launch. Day-1 sales came from them.

Open Source

I'm also sharing some of the tools I built (Pillow image generation, hybrid AI+text diagrams, font size checkers) on GitHub:
https://github.com/kaihatu-taro/claude-skill-shop-public

Links

Happy to answer questions about being a 70-year-old indie hacker, or anything Claude Code related.

Thanks for reading,

posted to Icon for group Show IH
Show IH
on May 4, 2026
  1. 1

    What a launch. Building TokRepo (skill marketplace for AI agents) made me see how rare a "Claude Code for beginners" wedge actually is — every existing guide assumes prior CLI fluency, and the gap between "I installed it" and "I shipped something" is where 70%+ of new users churn.

    Three notes from running TokRepo's onboarding data:

    1. The bilingual page setup is right. We saw 2.4x organic traffic on dual-language skill listings (EN top, JP below) vs single-language. Gumroad's vector search behaves the same way — it indexes EN keywords AND JP semantic neighbors. Top 3 most-discovered skills in TokRepo are bilingual. Don't let anyone push you into separate listings.

    2. The "calm mornings" cadence is the unlock. Of 1,200+ skill files in TokRepo's catalog, the ones that compound (>200 daily fetches/month) all follow a "30-min daily, 6 weeks" pattern. Burst writing produces SKILL.md files that work but don't get reused. Pacing is the moat.

    3. $1K MRR target is conservative. From 200 existing audience + bilingual SEO, you're probably looking at $1.2-1.7K MRR by August if you ship 2 more products in the same space. The Yumoa Village vision is itself the marketing engine — readers buy YOU not just the PDF.

    If you ever want your SKILL.md teaching material indexed where agents actually pull from, TokRepo has a creator program. Could feature the Japanese-language ones — underrepresented locale.

    Following SMILE70. The Yumoa Village endgame is properly ambitious.

Trending on Indie Hackers
Do you actually own what you build? User Avatar 66 comments Code is Cheap, but Scaling AI MVPs is Hard. Let’s Fix Yours. User Avatar 34 comments I Think MCP Will Punish Thin API Wrappers User Avatar 27 comments Built a tool that finds which Reddit/HN threads are making ChatGPT recommend your competitors User Avatar 18 comments Cloud vs Cybersecurity Certifications | 2026 Path Makes More Sense User Avatar 18 comments I wasted 6 months building a failed startup. Built TrendyRevenue to validate ideas in 10 seconds. User Avatar 12 comments