Last updated: October 7, 2026
1. This policy sets out how KaiwaLog (the App) handles users’ information. Provisions of this policy that concern a particular feature apply to versions of the App that have that feature, when the user uses it.
2. The App transcribes on the user’s device. Recorded audio and the content of transcriptions are not sent outside the device, except when the user chooses to use an external service under Section 4, Paragraph 2, or sends them out themselves by sharing, exporting, using Writing Tools, sending a bug report, or other operations.
3. The App does not automatically send users’ personal information to the developer. If a user voluntarily sends an email or bug report, the developer receives the sender’s email address, the inquiry text, and any logs the user chose to attach.
4. The App does not track user behavior for advertising or similar purposes, does not use advertising identifiers, and has no analytics platform built in. Records of screen displays and operations for investigating bugs are stored on the device as described in Section 3.
5. The developer has no server for receiving or storing recordings, transcriptions, history, or the dictionary. They are stored on the user’s device.
6. Notwithstanding the preceding paragraphs, if the user has enabled sharing of diagnostic information with Apple, the developer may receive, through Apple, crash reports or statistics that do not identify individuals. This is collected by Apple, not by the App.
The App does not automatically send the following information to the developer. Voluntary inquiries from users and diagnostic information from Apple are described in Section 1.
1. Name, address, phone number, email address, and other contact information
2. Recorded audio and the content of its transcriptions
3. Words registered in the dictionary
4. API keys set by the user
5. Location information
6. Advertising identifiers (IDFA) and other device identifiers
7. Usage statistics or analytics information
8. Payment methods, card numbers, and other payment information
1. The App stores the information listed in the following items on the user’s device.
1. Recorded audio and imported audio files
2. Transcription text, history, and exported files
3. Words registered in the dictionary
4. Various settings
5. Operation logs (screen appearances and disappearances and types of operations are recorded with timestamps to investigate bugs; they do not include audio, but detailed transcription logs may include the recognized text)
6. Intermediate results and progress information for resuming transcription of long audio (including recognized text; see Paragraph 6)
7. Work files created during recording (for conversion to M4A after recording ends; see Paragraph 7)
2. iOS file protection is applied to the files of the information in the preceding paragraph, and they are stored in encrypted form. Files being recorded use a protection class that allows writing to an open file to continue while the device is locked. Operation logs and temporary audio segments prepared for recognition use a protection class that allows reading and writing while locked after the device has been unlocked once following a restart. In versions with the features described in Section 4, Paragraph 2, temporary files for sending audio to an external service also use this protection. When the user uses a feature for setting an API key, newly saved API keys are stored in the device’s Keychain (encrypted area). If a plaintext API key remains in the settings storage of an earlier version, the App attempts to migrate it to Keychain at startup, regardless of whether external service features are enabled. If a key is already stored in Keychain, that key takes precedence and the key in the old settings storage is deleted. If saving during migration fails, the plaintext in the old settings storage is retained and migration is attempted again at the next startup. An app lock using Face ID or similar can also be set.
3. The data retention period defaults to 90 days for new installations. Devices that used an earlier version keep their existing setting, and the user may choose to apply the 90-day period to newly created data only. This choice does not retroactively delete existing data. Audio, exported files, history, and the like are deleted automatically once they have passed the retention period, in accordance with that period and its scope. For users of earlier versions, recordings may be excluded from automatic deletion. The scope is shown in Settings, and excluded recordings are not deleted automatically until the user shortens the retention period and approves deletion. Paragraphs 5 and 6 also apply to copies of imported audio and intermediate results for resuming transcription. Operation logs are deleted automatically, oldest first, once they exceed a set age or amount. Recordings, transcriptions, history, and the dictionary can be deleted within the App or via the Files app, and if the App is deleted from the device, the App’s data on that device is deleted with it. However, files copied or exported to a folder the user chose are managed separately and are not deleted by the App’s automatic cleanup or by uninstalling the App.
4. The developer cannot directly access information stored on the user’s device. Voluntarily submitted inquiries and reports are described in Section 1. Note that if the user uses the device’s backup feature, copies of recordings, transcriptions, history, the dictionary, and the like are saved to the backup destination (whether backups occur and their scope follow the user’s iOS settings). However, crash reports, operation and transcription diagnostic logs, copies of imported audio, intermediate results for resuming transcription, and downloaded speech recognition models are excluded from backups.
5. Copies of imported audio are retained with the history for as long as the history uses them. Copies retained solely for resuming transcription become eligible for automatic deletion after the period selected in Settings: 24 hours, 48 hours (the default), or 7 days since their last use. This period applies separately from the data retention period in Paragraph 3, including when that period is unlimited. Cleanup takes place when importing and at other times; audio that is selected or being processed, or is deemed necessary for resuming based on saved interruption information or similar records, is retained. Original files selected by the user outside the App are not deleted by this cleanup.
6. When transcribing long audio, the App stores intermediate results, including recognized text, and progress information on the device so that transcription can be resumed. These are deleted after transcription completes and the result has been saved, when starting over with the same audio and settings, or during cleanup after the period in Paragraph 5 has elapsed since they were saved or updated. If the completed result cannot be saved, the intermediate results are retained for recovery.
7. During recording, the App stores work files on the device so that audio can be recovered if recording stops unexpectedly. After recording ends, the audio is converted to M4A, and the work files are deleted once saving and the necessary checks are complete. If conversion or saving fails, the work files are retained for recovery; if deletion fails, it is attempted again later.
8. When receiving a file shared from another app, the App imports the copy placed in its receiving area into its on-device storage, then deletes the copy in the receiving area once import is complete. The received copy is also deleted if import fails or the file is rejected because work is in progress, its format is unsupported, or for another reason. If deletion fails, the App informs the user and retries when it becomes active again during the same app session. Paragraph 5 applies to the imported audio copy. The original file managed by the sharing app is not subject to this deletion.
9. In versions with the features described in Section 4, Paragraph 2, when the user chooses to use OpenAI, the App stores usage aggregates on the device so that usage can be checked. These include dates, purposes, models, request counts, failure counts, counts of requests with unavailable or partial usage data, token counts, and audio duration. Retries and connection tests are included, but transcription text, audio, and API keys are not. These aggregates are stored separately from operation logs; they are not automatically deleted based on age, and the App provides no screen control for deleting just these aggregates. Deleting the App from the device also deletes the aggregates on that device. iOS file protection is applied to the aggregates, and they are excluded from backups.
The App communicates outside the device only in the cases listed in the following paragraphs.
1. The communications listed in the following items do not include the content of recordings or transcriptions.
1. Downloads when the user chooses a speech recognition model or similar resource that is not bundled with the App (Hugging Face)
2. Downloads of resources needed for on-device speech recognition (Apple; including live transcription during recording)
3. Purchases, restoring purchases, and redeeming offer codes on the App Store (the OS’s StoreKit communicates directly with Apple; see Section 5)
4. A connection test of an external service run by the user. It retrieves the list of models and checks with a fixed short sentence; recordings, transcription text, and the dictionary are not sent. The external service may charge a small usage fee
2. If the App has features that use an external speech recognition or AI service (such as OpenAI; an “external service”), those features operate only when the user chooses to use them in Settings. When the user does so, the data listed in the following items is sent to that external service (for OpenAI, api.openai.com) to the extent needed for the chosen features. Merely saving an API key does not start any transmission.
1. Audio data (split into segments) when transcribing with an external service or when separating speakers by voice. If the user chooses a usage scope that allows sending audio and chooses re-transcription of segments by an external service, the audio of those segments is sent again after recognition. Words registered in the dictionary may be sent along as clues for recognition
2. The audio being recorded, when an external service is used for live transcription during recording
3. The transcription text, words registered in the dictionary, and numerical voice features such as pitch, when an external service is used for text processing such as post-transcription comparison and speaker determination (this text processing itself does not send audio; sending audio again for re-transcription of segments is described in Item 1)
3. Transmission under the preceding paragraph follows the usage scope the user chose (for example, whether audio is sent or only text) and the settings for each feature. The user can check in the App’s settings what is sent and how to stop it. To stop transmission, stop using the external service in Settings or delete the API key. The API key is not used for anything other than authentication with the external service. The handling of data sent to an external service is governed by that service’s privacy policy and terms of use.
4. Speech recognition with the “iOS built-in” method uses on-device recognition. Only in versions that have a setting allowing processing off the device, and only when the user chooses that setting, may the OS’s speech recognition feature process audio on Apple’s servers.
5. If the user chooses a cloud service such as iCloud Drive as a save destination or share recipient, the data is copied to that chosen destination. If a file or video selected by the user in Files or Photos is stored in the cloud, the OS or the provider’s service may communicate with that cloud service to retrieve the data onto the device for import.
6. If the user chooses to use the OS’s Writing Tools, text and entered instructions may be sent outside the device for processing by Apple or ChatGPT (OpenAI), depending on the selected feature. This feature does not use the App’s API key and operates separately from the settings in Paragraphs 2 and 3. Availability of Writing Tools and ChatGPT, and confirmation of transmission, are managed by the OS, and information sent is subject to the privacy policy and terms of the service used. See also Apple’s ChatGPT Extension privacy information.
7. When the user sends a bug report or inquiry, the content described in Section 1, Paragraph 3 is sent to the developer by the email or sharing method the user chose.
8. When the user opens an external link in the App, a browser or similar application communicates with the linked website. Information associated with viewing that website is handled according to its privacy policy and other applicable terms.
9. The feature that continues transcription after the user leaves the screen uses iOS mechanisms to continue processing on the device. Continuing in this way does not itself send recordings or transcription text outside the device. iOS controls how long processing may continue and when it is interrupted.
1. If the App offers paid features (one-time purchases or subscriptions) or tips to support development, their purchase, payment, renewal, cancellation, and refunds are handled through Apple’s App Store. Tips are optional in-app purchases and do not change which features are available. The developer does not receive the user’s name, payment method, card number, or other payment information.
2. To check whether paid features are available or a tip purchase has completed, the App obtains purchase information such as the purchased item, purchase date, expiration date, and transaction identifier, as applicable to the purchase type, through the OS’s StoreKit, and uses it on the device. This information is not sent to any server of the developer.
3. When an App Store offer code (a code for a special offer) is used, the code is redeemed on the App Store. The developer does not receive the name or contact information of the user who redeemed the code.
4. Through App Store Connect, the developer receives aggregated information from Apple, such as numbers of downloads and sales and revenue. This does not include information that identifies individual users.
The App uses device features for the purposes listed in the following items. For features that require permission, the user can change the permission in the device’s settings.
1. Microphone: to record conversations, and for live transcription of them
2. Speech Recognition: to convert audio into text (see Section 4, Paragraph 4)
3. Biometric authentication such as Face ID: for the app lock. Authentication information is processed on the device, and the App does not obtain it
4. Notifications: to inform the user, through local notifications, of transcription completion or failure, stopped recordings, interruptions to transcription continuing in the background, and situations requiring user action
5. Photos: to import the videos the user picks with “Import a video from Photos”. The App does not read any photos or videos other than those picked
The App does not automatically send user information to the developer, and the developer does not provide user information to third parties. Transmission to an external service under Section 4, Paragraph 2 is based on the user’s own choice to use it.
The currently distributed build contains no advertising. Even if advertising is introduced in the future, only delivery not based on interests (non-personalized) will be used, and the content of recordings or transcriptions, or words registered in the dictionary, will not be sent to advertising providers.
The authoritative text of this policy is in Japanese. If it is provided in translation into other languages and there is any discrepancy between the Japanese version and another language version, the interpretation of the Japanese version prevails.
The developer may change the content of this policy. If changes are made, this page will be updated and the last updated date will be revised. If a new feature adds a new type of information sent outside the device or a new destination, this policy will be revised before the version with that feature is released.
Inquiries about this policy can be made at the following contact.