Keep the destination in Japanese
Save the venue’s own Japanese name and address, with a map pin where available. A romanized name helps an English reader, but it should not replace the original text. Label each version so the other person can choose the useful one. Do not invent characters for a personal name or try to repair an unfamiliar place name from memory.
Fictional exercise: prepare an address card with an original venue name, building number and entrance note. Check whether the translation tool preserves the destination when you translate a request around it. Ask a fluent reviewer whether the meaning is intact; matching English on a back-translation is not enough.
Test a request, its subject and the reply
| Test case in English | What the reviewer should check | What you can check visually |
|---|---|---|
| I have a reservation for two people. | The speaker and request remain clear. | The numeral or count stays two. |
| The other person asks for the booking name. | The reply is a question, not a confirmed booking. | The original name is copied correctly. |
| The time changes from 3 to 4. | The correction replaces the old detail. | Only the final agreed time is marked current. |
These are English test prompts, not published Japanese translations. Supply the conversational context instead of assuming an isolated sentence gives the system enough information about who is acting.
Practise a slower reply and a repeat
Japan Foundation’s Irodori lesson explicitly practises asking someone to repeat and to speak more slowly. The finite Japanese phrase sheet includes those reference-checked lines with script, English meaning and romanization. They have AI/editorial review, not native-speaker certification or recorded audio.
Keep a number written down even after a repeat. An acknowledgement does not itself prove agreement. For a spoken-output test, ask the reviewer to assess what they heard without first reading the English answer.
Record the mode that actually worked
Log typed input, image input, microphone recognition and spoken output independently. A product may expose different controls for each; this guide does not establish their availability. For choosing what to test, use the Japanese app selection checklist.
Sources and testing scope
This is an app-agnostic preparation checklist. The practice cases are editorial exercises, not measured app output, a feature availability claim or a native-speaker review. Check any product’s current release and the exact language direction before using it.
Use the app test sheet to record results and the offline readiness planner for connection checks. Pocket’s current public product information is on Support.