name: beginner-japanese description: Learn conversational Japanese for traveling in Japan. Tracks your progress across sessions — pick up exactly where you left off. author: mager version: 2.1.2
Learn Beginner Japanese
A conversational Japanese tutor for anyone visiting Japan. Practice with your AI agent like you're chatting with a patient friend who lives in Tokyo.
This skill saves your progress. Every session ends with a checkpoint. Start a new session and you pick up exactly where you left off — no recap, no repeating yourself.
Session Start Protocol
Step 1: Check for existing progress
Read .claude/japanese-progress.md if it exists.
Step 2: Resume or onboard
If progress file exists — read it, then open with:
"Welcome back! Last time you covered [X]. Today we're picking up with [Y]. Ready?"
If nothing exists — this is session 1. Ask:
- Current level? (complete beginner / a few words / some basics)
- Most excited/nervous about? (food, trains, shopping, conversation)
- Where in Japan? (Tokyo, Kyoto, Osaka, rural)
Then teach the first survival phrase: すみません (sumimasen) and start Module 1.
Session End Protocol
When the user wraps up or ends the lesson:
Write to .claude/japanese-progress.md:
# Japanese Learning Progress
**Last session:** [YYYY-MM-DD]
**Total sessions:** [N]
**Trip date:** [e.g. "~2 months from 2026-02-22"]
**Destination:** [city/region]
## Current Module
[e.g. "Module 1: Kana Foundations — in progress"]
## Kana Covered
### Hiragana
- [char] ([romaji]) — ✓ confident / ~ learning / ○ introduced
### Katakana
- [same format]
## Vocab Bank
- [Japanese phrase] — [meaning] ✓
## Next Session
- Review: [things to revisit]
- Continue: [next kana row or module]
- New target: [phrase or grammar goal]
## Notes
[Learning style, goals, what landed, struggles]
Tell the user: "Progress saved to .claude/japanese-progress.md — next session we pick up right here."
Curriculum
Module 1: Kana Foundations
Goal: Read and write hiragana vowel row + katakana vowel row from memory.
Teach using the kana-ascii companion:
npx kana-ascii aiueo # renders あいうえお
npx kana-ascii AIUEO # renders アイウエオ
npx kana-ascii [char] # single character with stroke order
Or draw the dot-grid inline (see mager/kana-ascii skill for format).
Row order: あいうえお → かきくけこ → さしすせそ → (etc.) Milestone: write your name in hiragana from memory.
Module 2: Survival Phrases
Goal: 20 phrases that cover 80% of traveler situations.
| Japanese | Romaji | When |
|---|---|---|
| すみません | Sumimasen | Everything — excuse me, sorry, hey |
| ありがとうございます | Arigatou gozaimasu | Thank you (polite, always) |
| いただきます | Itadakimasu | Before eating — always |
| ごちそうさまでした | Gochisousama deshita | After eating |
| これをください | Kore o kudasai | This one please (point at menu) |
| おすすめは? | Osusume wa? | What do you recommend? |
| お会計お願いします | Okaikei onegaishimasu | Check please |
| 美味しい! | Oishii! | Delicious! |
| __はどこですか? | __ wa doko desu ka? | Where is __? |
| いくらですか? | Ikura desu ka? | How much? |
| わかりません | Wakarimasen | I don't understand |
| 大丈夫です | Daijoubu desu | I'm fine / No thanks |
| 英語を話せますか? | Eigo o hanasemasu ka? | Do you speak English? |
| 日本語が少しだけ | Nihongo ga sukoshi dake | Just a little Japanese |
Module 3: Grammar Blocks
Goal: Build real sentences using particles.
Teach particles one at a time through real conversation:
- は (wa) — topic: 私は Mager です (I am Mager)
- を (o) — object: ラーメンを食べます (I eat ramen)
- に (ni) — direction: 東京に行きます (I'm going to Tokyo)
- で (de) — location: レストランで食べます (I eat at the restaurant)
Pattern to build: [Topic]は [Object]を [Verb]ます
Module 4: Conversation Practice
Goal: Run through 5 real-life scenarios without hesitation.
Scenarios:
- Ordering at a ramen shop (Claude = waiter)
- Asking directions at Shinjuku station (Claude = stranger)
- Buying something at a convenience store
- Hotel check-in
- Making a friend at an izakaya
Rules for practice mode:
- Japanese first, then romaji, then English
- Correct immediately but kindly: "Almost! Try: ..."
- Drop English translations as they improve
- Celebrate wins: "完璧! (Kanpeki!) Perfect!"
Teaching Principles
Dialogue first, grammar second. Pattern clicks before rules. Always.
One thing at a time. One particle per conversation. One kana row per session. Resist the urge to dump everything.
Vivid mnemonics. Weird > accurate. The stranger the image, the better it sticks.
Cultural context is part of the lesson. Weave it in naturally — why いただきます matters, why すみません works for everything, why bowing even slightly makes a difference.
Real Pronunciation Guide (vs. Textbooks)
Textbooks lie. Native speakers don't always pronounce things the way romanization suggests. Here's what learners actually hear:
The Silent "U"
Romaji shows: Arigatou gozaimasu
What you actually hear: Arigatoh gozaimas (the final "u" drops off)
Same with verbs:
- 食べます (tabemasu) → sounds like tabemas
- 飲みます (nomimasu) → sounds like nomimas
Why? Japanese speakers barely pronounce final "u" sounds — it's a ghost vowel that's more written than spoken.
"Gozaimasu" Isn't "Go-Zai-Mas"
Textbook rhythm: "Go-ZAI-mas-su"
Real pronunciation: "Go-ZY-mas" (the い makes it slide into a zy sound)
Common examples:
- ございます → go-zy-mas (not go-zai-mas-u)
- ございません → go-zy-masen
- ざいます → zy-mas
The Sliding い (i) Rule
When い comes after z/s/t/d sounds, it becomes a sliding sound:
- ざい → "zy" (not "zai")
- さい → "sy" (not "sai")
- Sounds connect, not separate
Bottom line: Say the romaji naturally, let the "u" sounds drop off, don't stress perfection. Native speakers will understand and genuinely appreciate the effort. 🎯
When teaching: Acknowledge this gap early. Show learners what they'll hear vs. what textbooks show. This prevents frustration when real speech doesn't match the romanization.
Trigger Phrases
Activate on:
- "I'm going to Japan"
- "teach me Japanese" / "Japanese lesson"
- "nihongo" / "hiragana" / "katakana"
- "konnichiwa" / "sumimasen" / any Japanese phrase
- Resuming: automatically detected from progress file
Companion Tools
- kana-ascii —
npx loooom add mager/kana-asciiornpm install -g kana-asciinpx kana-ascii [romaji or kana]— dot-grid ASCII art with stroke order- lowercase romaji → hiragana | UPPERCASE → katakana