Project Emily is an AI-assisted JLPT N4 study app based only on the TRY! 日本語能力試験 N4 word bank.
Current MVP:
- Google sign-in
- chapter-based kanji and vocabulary exams
- saved exam history
- progress tracking by chapter, entry, and question pattern
- PROJECT_EMILY_MVP_PRD.md
- docs/content-model.md
- docs/content-caveats.md
- docs/db-design.md
- db/project_emily_schema.sql
- Treat
content/try_n4/as the source of truth for authored content. - Treat
build/as generated import artifacts, not authoring files. - Keep
entry_idstable across lexicon, POS, examples, and distractors. - Use the database for runtime reads and user progress, not raw CSV files in production.
- For question generation, the database decides the correct answer. The model only packages the question.
usagequestions require more review thankanji_reading,orthography, orparaphrase.
.agents/skills/project-emily-content-workflow/SKILL.md.agents/skills/project-emily-question-generation/SKILL.md.agents/skills/project-emily-mvp-implementation/SKILL.md
These skills are repo-local guides for Codex work on this codebase. If you want them to auto-appear in Codex's global skill list, copy or install them into your Codex skills directory later.