Privacy Policy
trisaveurs ("we", "us") sets out below its Privacy Policy (this "Policy") concerning how we handle users' personal information in "Exevoice" (the "Service") that we provide.
Article 1 (Personal information)
"Personal information" means personal information as defined in the Act on the Protection of Personal Information: information about a living individual that can identify that specific individual by the name, date of birth, address, telephone number, contact details or other description contained in it, together with information that can identify a specific individual on its own (personal identification codes), such as data on facial appearance, fingerprints and voiceprints, and the insurer number on a health insurance card.
Article 2 (Information we collect and how we collect it)
In providing the Service, we collect the following information.
Information stored on our servers
- Email address (via Google account authentication)
- Session record text (structured text organized by AI from your recorded audio: summary, your words, physical, memory, share text and so on)
- Full session transcript (the full text after the transcription of your audio has been corrected for notation, and before AI summarizes and organizes it. It is not shown in the app screens and is used only as an internal reference when generating features that look back and observe across your accumulated records)
- Session date and recording length
- Purchase history and subscription status for users on paid plans (obtained via RevenueCat)
- Record of consent to sending data to outside AI services (the date and time of consent, and the version of the disclosure shown at that time)
Information collected automatically as you use the Service
- Device information (OS version, device type and so on)
- Usage logs (access date and time, feature usage and so on)
- Error information (crash reports and so on)
Information we do not store on our servers
The Service does not store on our servers the audio you record, or the transcribed text before notation correction. These are sent temporarily to outside AI processing services and are discarded once processing finishes. See Article 4 for details.
The Service records what you choose to say, and what you say may include references to bodily sensations or how you are feeling physically. We do not actively collect or analyse these as information for medical or health management purposes; we treat them as part of your own journal for looking back.
Article 3 (Purposes for collecting and using personal information)
We collect and use personal information for the following purposes.
- To provide and operate the Service
- To provide features that look back and observe across your past session records
- To respond to your enquiries
- To send email about new features, updates and campaigns for the service you use, and about other services we provide
- To contact you as necessary, for maintenance and important notices
- To identify users who breach the Terms of Service or who try to use the Service for improper or unjust purposes, and to refuse them use
- To let you view, change or delete your registered information and view your usage
- For purposes incidental to the above
Article 4 (Handling of audio data and what you say)
Prior consent to sending data to outside AI services
Before sending your data to the outside AI services described below, we disclose in a screen inside the app the types of data sent, the names of the companies they are sent to, and the purpose of sending, and we obtain your consent.
If you do not consent, no recording or transcription takes place and no data is sent to outside AI services at all. Even then, you can still read the session records already stored, as before.
You can review what was disclosed at any time from "About AI processing" in the app's settings screen. The record of consent (the date and time, and the version of the disclosure) is stored on our servers.
Audio and pre-correction text are not stored
The Service does not store on our servers the audio you record or the transcribed text before notation correction. They are processed as follows.
1. The recorded audio is sent through our servers to OpenAI's Whisper API, which produces transcribed text. The audio is never stored on our servers.
2. The transcribed text is corrected for notation using a dictionary we maintain, then sent to Anthropic's Claude API, where it is organized into structured text (the "session record"): summary, your words, physical, memory, share text and so on.
3. Once the session record has been generated, the audio and the transcribed text before notation correction are discarded.
Storage of session records and full transcripts
Session records are stored in the Asia Pacific (Tokyo) region of Supabase, in a form that only you can read.
The full transcript after notation correction is stored in the same region, linked to your account. This full transcript is not shown in the app screens and is used only as an internal reference when generating features that look back and observe across your accumulated records. We do not actively read or analyse individual content.
About training use at outside AI services
When we use the OpenAI and Anthropic APIs, we send your input with settings that exclude it from training their machine learning models. Please refer to each company's API usage policies for details.
About references to your body and physical condition
What you choose to say may include descriptions of your body: its state, sensations, discomfort and so on. We state the following about such descriptions.
- We do not provide medical advice, diagnosis or treatment suggestions through the Service. Text organized by AI is your own speech structured so that it is easier to read back; it is not intended to be used as a basis for medical judgement.
- The Service does not connect with Apple HealthKit or any other health data platform, and does not obtain health data such as heart rate, step count or sleep.
- What you say is stored as your own journal for looking back, and we do not actively analyse or profile it.
- If you need a basis for your own judgement about poor condition, pain or symptoms, please always consult a medical institution.
About the share feature
If you choose to publish content to an outside service (such as X) using the share feature, that content becomes public information and is handled in accordance with the privacy policy of that outside service. Managing or deleting it after sharing is your own responsibility.
Article 5 (Outside services and transfer of information overseas)
We use the following outside services in providing the Service. In these services, your information may be stored and processed on servers located in the country where the service is based.
Supabase Inc. (United States)
Database and authentication. The data storage region is Asia Pacific (Tokyo).
Privacy policy: https://supabase.com/privacy
RevenueCat, Inc. (United States)
Managing purchases of paid plans (auto-renewing subscriptions) and checking subscription status. We do not obtain payment information such as credit card numbers.
Privacy policy: https://www.revenuecat.com/privacy/
OpenAI, L.L.C. (United States)
Transcribing audio (Whisper API). We send data with settings that exclude it from training.
Privacy policy: https://openai.com/policies/privacy-policy
Anthropic, PBC (United States)
Organizing transcribed text (Claude API). We send data with settings that exclude it from training.
Privacy policy: https://www.anthropic.com/privacy
Google LLC (United States)
Authentication service (Google login).
Privacy policy: https://policies.google.com/privacy
Mixpanel, Inc. (United States)
Analysis of how the Service is used.
Privacy policy: https://mixpanel.com/legal/privacy-policy/
Sentry (Functional Software, Inc.) (United States)
Recording and analysing errors.
Privacy policy: https://sentry.io/privacy/
We use these companies having confirmed that they provide a level of protection for personal information equal to or higher than this Policy. By using the Service, you are deemed to have consented to the transfer of information overseas described above.
However, for the sending of audio and transcribed text to OpenAI and Anthropic, separately from this blanket consent arising from use, we disclose what is sent inside the app and obtain your individual consent, as set out in Article 4.
Article 5-2 (Handling of payment information)
Purchase, payment, billing, refunds and cancellation for the paid plans of the Service all take place through the App Store mechanisms provided by Apple Inc. We do not obtain any information about your credit card number, bank account or other payment methods.
What we obtain is limited to the following information, by way of Apple and RevenueCat.
- The name of the plan you are subscribed to
- The contract start date, renewal date and end date
- The identifier of the purchase transaction
We use this information only to determine whether you are in a state where you can use the features of a paid plan.
For details of paid plan fees, how to cancel, refunds and so on, please also see Article 10 of the Terms of Service.
Article 6 (Provision of personal information to third parties)
Except in the following cases, we do not provide personal information to third parties without obtaining your consent in advance.
- Where required by law
- Where it is necessary to protect a person's life, body or property and it is difficult to obtain the person's consent
- Where it is particularly necessary to improve public health or promote the sound growth of children and it is difficult to obtain the person's consent
- Where it is necessary to cooperate with a national government body, a local government, or a party commissioned by them, in carrying out duties prescribed by law, and obtaining the person's consent risks impeding those duties
Article 7 (Data retention period)
We retain your account information, session records and full transcripts for as long as you hold an account.
If you delete your account from the "Settings" screen in the Service, we immediately delete from our database your account information, session records, full transcripts, usage records and your record of consent to sending data to outside AI services. Deleted data cannot be restored.
However, we may retain the following information to the extent necessary.
- Information whose retention is required by law
- Information anonymised as statistical data
Note that information already provided to the outside services listed in Article 5 may continue to be retained by those companies in accordance with their own policies and retention periods, even after deletion from our database.
Note also that the period during which you can access past session records on our Service may differ depending on your plan (free or paid). Records beyond the accessible period are still retained as data for as long as you hold an account.
Article 8 (Disclosure, correction and suspension of use of personal information)
You may ask us to disclose, correct, suspend the use of, or delete your personal information.
Deleting your account and the data linked to it: you can do this yourself at any time by choosing "Delete account" from the "Settings" screen in the Service. Handling is as set out in Article 7.
For other requests (disclosure, correction, suspension of use), please contact us by email at the address below.
Contact: exevoice.app@gmail.com
We will respond within a reasonable period after confirming that the request comes from you.
Article 9 (Changes to this Privacy Policy)
Except for matters prescribed by law or otherwise stated in this Policy, the content of this Policy may be changed without notice to you.
Except where we separately provide otherwise, a changed Privacy Policy takes effect when it is posted on the Service.
Article 10 (Contact)
For enquiries about this Policy, please contact us at the address below.
trisaveurs / Contact: exevoice.app@gmail.com
Article 11 (Language)
The Japanese text of this Policy is the authoritative version. The English version is a translation provided for reference; if there is any discrepancy between the Japanese and English versions, the Japanese version prevails.