Kanji Shisho

The description states that Kanji Shisho teaches Japanese reading — including hiragana, katakana, and 1,026 MEXT Grade 1–6 kanji — through verified study cards, practice sessions, and applied sign-reading missions. It features a resident guide named "Shisho" (司書), which lives in the header of every page.

The system uses verified dictionary data (MEXT, KANJIDIC2/JMdict, KanjiVG) for facts and logic, while GPT-5.6 is used only for explanations. The model receives a minimal fact envelope; it does not generate answers but explains mistakes live in the learner's language.

The product is built with AWS Lambda, API Gateway, CloudFormation, S3, CloudFront, Node.js, TypeScript, and Codex. It includes CI/CD pipelines, infrastructure-as-code, and security scanning.

Inference The product is an AI-enhanced educational platform focused on Japanese reading comprehension, using a hybrid of deterministic logic and generative AI for personalized feedback.

The project positions itself as a third way between flashcard apps (which offer canned feedback) and chatbots (which may confidently invent readings). It claims to be a tutor that is “never wrong about facts” and always personal in explanation.

The author states:

• “We wanted a third thing: a tutor that is never wrong about facts and always personal in explanation.”
• “Verified dictionary data (MEXT, KANJIDIC2/JMdict, KanjiVG) and deterministic logic own every fact, reading, score, and answer. GPT-5.6 owns only the explanation.”

This positioning emphasizes trust, accuracy, and personalization.

Inference The product is positioned as an educational tool that balances AI personalization with factual integrity — a niche in the language-learning space.

The description states that Kanji Shisho targets learners of Japanese reading, specifically those studying hiragana, katakana, and MEXT Grade 1–6 kanji. It is designed for users who want to progress through structured lessons and receive accurate, personalized feedback.

Inference The target customer is likely Japanese language learners at beginner to intermediate levels, particularly those using formal curricula (e.g., MEXT standards).

There is no evidence of pricing or business model in the description. The project is presented as a hackathon submission with no mention of monetization, subscriptions, or paid features.

Inference The business model is not evident. It may be early-stage or non-existent at this point.

The system is built using serverless AWS infrastructure (Lambda, API Gateway, CloudFormation), with CI/CD pipelines using Codex. The team used deterministic quality gates including 200+ tests, 8,963 exercise contracts, and full-tree security scans.

Key technical details:

• Uses GPT-5.6 for explanations only.
• Output is capped at 192 tokens to meet latency requirements.
• Implements fail-closed guards to prevent model drift.
• Infrastructure is managed with CloudFormation and reviewed via change sets.
• CI/CD pipeline runs in ~2m38s per run.

Inference The technical stack is robust for a hackathon-scale product and shows attention to security, reproducibility, and performance.

There is no evidence of user traction, adoption, or revenue. The project is described as a hackathon submission with no mention of users, customers, or usage metrics.

Inference The product is at a very early stage and lacks any demonstrated traction or maturity in the market.

There are no references to competitors or market positioning beyond the claim that this is a “third thing” between flashcards and chatbots. The author does not name or describe existing tools in the Japanese language-learning space.

Inference The competitive context is unclear. It may be competing with educational apps and AI-powered tutors in the Japanese language-learning space, but no direct comparison is made.

1. No traction or revenue: The project is described as a hackathon submission with no evidence of adoption.
2. Unverified claims: All claims are self-reported and unverified.
3. Single founder: The team size is listed as 1, which may limit scalability or execution capacity.
4. AI dependency: Reliance on GPT-5.6 for explanations introduces risk if the model becomes unavailable or misaligned.
5. Limited scope: The product focuses only on MEXT Grade 1–6 kanji and hiragana/katakana — a narrow educational domain.

Inference The project is at a very early stage and faces risks related to scalability, execution, and market validation.

1. What is your plan for scaling beyond a single developer?
2. How do you intend to monetize this product?
3. Have you validated demand with potential users or educators?
4. What are the risks of relying on GPT-5.6 for explanations, and how do you mitigate them?
5. Are there any plans to expand beyond MEXT Grade 1–6 kanji?
6. How do you intend to build trust in the educational space without a track record?