Mojiokoshi (the “App”) performs all speech-to-text processing entirely on your device (offline). The App is designed with your privacy as the top priority.
Information We Collect
The App does not collect, transmit, or store any user data, including personally identifiable information.
Audio & video: Recordings and selected videos are processed on-device solely for transcription; temporary files are deleted afterward. They are never sent off the device.
Transcripts (history): Generated text is stored only on your device. The developer cannot access it.
Analytics & tracking: No analytics, advertising SDKs, or tracking of any kind are used.
When Network Access Occurs
The App connects to the network only in the following cases, and never transmits your personal data:
Downloading recognition models: Speech-recognition models are fetched from Hugging Face (argmaxinc/whisperkit-coreml). Only model files are received.
Downloading translation language packs: The translation feature may download the language packs used by Apple’s on-device Translation from Apple. All translation happens on your device; your transcript text is never sent off the device.
Downloading the AI Advisor model: If you use High-Accuracy mode, a language model (Qwen) is fetched from Hugging Face (mlx-community). Only model files are received; your input text is processed on-device and is never sent externally.
Subscriptions: Purchases and restores are handled through Apple’s App Store / StoreKit; payment information is managed by Apple. Apple’s privacy policy applies.
Permissions
Microphone: Used to record audio for transcription.
Photo Library: Used to extract audio from a selected video for transcription.
Both are used only for on-device processing; captured audio/video is never sent externally.
Children’s Privacy
Because the App collects no user data, it is safe for all users, including children.
Changes to This Policy
This policy may be updated as needed. Significant changes will be announced on this page.