ARU (아름다운 루틴, 'beautiful routine' in Korean)

The description states that ARU is a mobile application that:

• Uses an on-device 30-second skin scan via MediaPipe Face Landmarker.
• Analyzes oil, redness, and texture from facial images.
• Provides a skin report, three product picks, and AM/PM routines in five languages.
• Includes optional email check-ins after two and four weeks.
• Supports full RTL layout for Arabic.
• Operates without requiring an account or API keys.

It also includes:

• A camera quality gate that checks light, distance, steadiness, and glare.
• An auto-capture mechanism triggered only when conditions are good.
• Integration with AI tools like GPT-5.6 for phrasing recommendations.
• Compliance features such as rate limiting, schema validation, and RLS on Supabase tables.

Inference The product appears to be a lightweight, self-contained PWA or native app built primarily using open-source technologies and AI assistants, with an emphasis on privacy (no data leaves the device) and accessibility (multi-language support).

The author claims ARU addresses a gap in existing K-beauty apps:

• Existing apps either over-promise efficacy ("whitened", "regenerated") or fail to answer the practical question: “What should I buy today?”
• ARU aims to provide honest, actionable skin insights and routines.
• It positions itself as a tool for both foreigners and locals who struggle with Korean ingredient labels.

Inference ARU’s positioning evolved from solving a personal problem (a model-turned-tech worker helping friends navigate K-beauty) into a scalable solution targeting global consumers seeking clarity in skincare choices. The shift toward AI-powered recommendation and localized content suggests an intent to expand beyond its initial use case.

The description states:

• ARU targets people visiting Korea who want to buy K-beauty products.
• It also serves Korean users who struggle with ingredient labels.
• Users are described as those who "can't read the labels" and need guidance in beauty stores.

Inference The primary customer segment appears to be international travelers or expats interested in K-beauty, with a secondary audience of local Koreans unfamiliar with product ingredients. The inclusion of five languages implies an intent to reach diverse global markets.

There is no evidence provided regarding:

• Revenue streams
• Pricing models
• Monetization strategy
• Customer acquisition costs
• Paid features or subscriptions

Not evidenced No mention of how the product will generate income, whether through direct sales, affiliate links, or other mechanisms.

The description indicates:

• Built using React, Next.js, TypeScript, Supabase, PostgreSQL, Playwright, Vitest, and MediaPipe.
• Uses Codex for feature development and GPT-5.6 for AI filtering of recommendations.
• Implements camera quality checks, auto-capture logic, and debugging overlays.
• Includes adversarial testing with Codex before launch.
• Full test coverage: 297 unit tests, 42 E2E mobile tests, and a text-fit matrix covering 180 combinations.

Inference The technical stack suggests a modern, well-tested product built for performance and scalability. The use of AI assistants in development implies rapid iteration capabilities, though the lack of production metrics raises questions about long-term viability.

The description states:

• 9 out of 10 moderated interviewees said they'd buy the exact product ARU picked.
• ~40 street interviews resulted in active engagement and waitlist signups.
• The app is live in production with full functionality (scan → survey → report → routine → check-in).
• The author plans to conduct a manual concierge round to convert stated intent into real transactions.

Not evidenced No data on actual conversions, revenue, or user retention. No mention of paid users, API usage, or marketing traction beyond the author’s own interviews and street testing.

The description states:

• Existing K-beauty apps either over-promise efficacy or fail to answer practical questions.
• ARU aims to be a more honest and actionable alternative.
• The app is positioned as a companion for people navigating Korean beauty stores.

Inference ARU competes in the skincare recommendation space, particularly where users are looking for clarity rather than marketing hype. However, there is no mention of competitors or market size — only self-perception of differentiation.

Key risks identified from the description:

• No verified traction or revenue: The product has not yet demonstrated real-world adoption or monetization.
• Unproven conversion rate: While users expressed intent, no data shows actual purchases or engagement beyond interviews.
• Dependence on AI tools: Heavy reliance on Codex and GPT raises concerns about scalability and control if these services change or become unavailable.
• Limited team size: Only one person is involved in the project, which may limit execution speed and depth of product development.
• Self-reported metrics: All evidence is self-reported; no independent verification exists.

1. What specific data do you have on conversion from waitlist signups to actual purchases?
2. How are you planning to scale beyond the current manual concierge approach?
3. Are there any plans for monetization or revenue models beyond affiliate links or direct sales?
4. How do you plan to maintain quality control as you expand into new markets or add more features?
5. What is your strategy for user retention and ongoing engagement post-initial scan?
6. Have you considered how compliance with international data privacy laws might affect the product?