English Japanese Talk — Privacy

Developer data collection

Audio, recognized text, and translations are not sent to developer servers or external AI services. There are no accounts, advertisements, tracking, or developer analytics SDKs.

Microphone and temporary audio

With permission, the microphone supports conversation translation and guide subtitles. Short PCM segments are processed on-device without saving audio files. Conversation holds a roughly 12-second utterance plus a short pre-roll. Guide mode bounds its queue to 8 segments / 45 seconds, plus the current recognition and capture segments of roughly 6 seconds each. Stopping or leaving a screen finishes or discards pending audio as described in the app.

History and preferences

Conversation history is limited to 30 in-memory entries and guide subtitles to 600 segments. They do not survive app termination. Clearing history requires confirmation. Preferences such as auto-start and voice speed use local UserDefaults; system backup follows Apple settings.

Apple language services

The app uses Apple's on-device speech recognition, translation, and speech synthesis. Language downloads require Apple network services. Apple may collect translation usage and performance information, excluding source and translated text. App Store and system diagnostics follow Apple's policies.

Consider other people

Explain that speech is being translated or captioned and obtain permission. Conversation contents appear on screen; consider shoulder-surfing and any screenshots you choose to share.

Support and external links

Opening support links contacts the website host. Emails you choose to send use your email address and message to answer your request. No app content is automatically attached. We update this policy when functionality changes. Developer: Koji Tamatsu.

Apple privacy policy · Apple Translation