KOIKUSE

KOIKUSE is a self-reported privacy-first relationship self-reflection web app built by one developer (Ra Ra) for personal relationship insight and practical next steps. The app presents 18 questions across six relationship dimensions, then offers one of four relationship tendencies with strengths, cautions, balance, and three actions. It also includes a 10-question check for another person’s behavior, presented as possibilities, not facts. The app uses local storage only, processes everything in the browser, and does not send data to servers. It was built using HTML, CSS, JavaScript, and OpenAI Codex (GPT-5.6). The author states it is a complete working product but provides no evidence of revenue, customers, or traction.

The description states that KOIKUSE is a privacy-first relationship self-reflection web app. It includes:

• 18 questions across six dimensions:
  • Security
  • Intimacy
  • Expression
  • Autonomy
  • Spark
  • Reassurance
• A diagnosis result with:
  • One of four relationship tendencies
  • Strengths and cautions
  • Relationship balance
  • Three practical actions
• A separate 10-question check for another person’s behavior, presented as possibilities, not facts
• Search functionality for local advice on crushes, messaging, dates, relationships, and breakups
• Results can be saved, reopened, and shared as image cards

The app is built with HTML, CSS, JavaScript, and uses OpenAI Codex (GPT-5.6) for development support.

Not evidenced: No information on how the 18 questions are scored or how the diagnosis logic works beyond the author’s description.

The author states that KOIKUSE is a tool that helps users understand themselves, organize observable behavior, and take respectful next steps without judging or claiming to know another person’s feelings. It is positioned as an alternative to tools that guess or manipulate.

It also claims to be:

• Privacy-first
• A complete working product (not a prototype)
• Built with AI coding tools (OpenAI Codex)

The app is described as a self-reflection tool, not a diagnostic or therapeutic one, and avoids presenting results as psychological or medical diagnoses.

Inference: The positioning suggests an emphasis on user agency and non-intrusive relationship guidance. However, the lack of evidence for adoption or feedback means this is unproven.

The description states that KOIKUSE helps users understand themselves and take respectful next steps in relationships. It includes a search function for advice related to crushes, messaging, dates, relationships, and breakups — suggesting it targets people navigating personal relationships.

It does not specify an age group or demographic beyond the general user of relationship tools.

Not evidenced: No information on who specifically uses the app, how many users there are, or whether it is targeted at a specific segment (e.g., teens, young adults, couples).

The description states that the app does not send diagnosis answers or searches to a server, and processing happens locally in the browser. It also says that only the latest diagnosis result is stored on the user’s device.

There is no mention of:

• Any monetization strategy
• Subscription or freemium model
• Paid features or premium content
• Pricing

Not evidenced: No evidence of a business model, pricing, or revenue streams.

The app was built with:

• HTML5, CSS3, JavaScript
• OpenAI Codex (GPT-5.6)
• Canvas API, Web Share API
• Responsive design

It is described as a complete working product, not a prototype.

Inference: The use of AI tools like Codex suggests that the developer may have used automation to build and iterate quickly. However, there is no evidence of scalability or long-term technical maintenance plans.

The author states that KOIKUSE is a complete working product, not a prototype. It was submitted to the OpenAI 2026 hackathon on Devpost.

There is no evidence of:

• User engagement or retention
• Customer acquisition or usage metrics
• Product iteration history
• Feedback from users

Not evidenced: No data on traction, adoption, or user behavior beyond the author’s self-report.

The description does not mention any competitors. It positions KOIKUSE as a tool that avoids guessing another person’s feelings or offering manipulative advice — which may differentiate it from some relationship apps or tools that attempt to infer emotional states.

Not evidenced: No information on existing tools in the space, their features, or how KOIKUSE compares.

• No evidence of user engagement or adoption: The app is described as a working prototype but lacks any traction data.
• Unproven effectiveness: There is no evidence that users find the self-reflection process or practical actions helpful.
• Limited scope: The app only includes 18 questions and 10 checks, with no indication of expansion plans or deeper content.
• Single developer: The team size is listed as one person (Ra Ra), which may limit scalability or long-term development.
• No monetization strategy: No evidence of a business model or revenue path.

Inference: Without user feedback or engagement, the app’s value proposition remains untested.

1. What is the expected user journey from first use to ongoing engagement?
2. How does the developer plan to validate that users find the self-reflection process useful?
3. Are there any plans for user testing or feedback collection?
4. How will the app evolve beyond its current 18 questions and 10 checks?
5. What is the long-term vision for monetization, if any?
6. How does the developer plan to scale beyond a single-person development model?