Privacy Policy

Effective: June 5, 2026 / Last revised: October 25, 2026 (storage of call transcripts at the host’s choice, temporary storage of imported files, the on-device temporary-storage cap, deletion/Trash/backups, proper-noun masking tables, on-device video adjustment, and the server-operation outsourcee clarified. Previous revision: September 3, 2026) / Provider: CoreVista-Japan, Inc.

This is a reference English translation. The authoritative version is the Japanese text; if there is any discrepancy, the Japanese version prevails.

CoreVista-Japan, Inc. (the “Company”) sets out the handling of personal information in the voice interpretation and records service “Cotoa Talk & Keep” (the “Service”) as follows.

1. Information We Collect

Audio is temporarily processed for translation, speech recognition, speech synthesis, or transcription and is not stored on the Company’s server beyond those processing purposes. The exceptions are: (1) the recording feature may temporarily hold unsent compressed audio on the user’s device to survive transmission failure or interruption; and (2) audio, video, and image files the user imports are kept in a temporary storage area on the Company’s server for transcription/text extraction and for retrying a failed job, for no longer than 24 hours after the job completes (7 days if the job fails or is cancelled), and are then deleted (Section 10). External providers may also temporarily process or retain data for abuse prevention, security, or incident response under their respective terms.

2. Purpose of Use

  1. Providing the Service: speech recognition, translation, caption display, speech synthesis
  2. Account management, balance management, recovery, billing, and payment processing
  3. Recording consent to the Terms and recording/audio processing
  4. Abuse prevention, security, incident response, quality improvement
  5. Responding to inquiries
  6. Compliance with laws and regulations
  7. Improvement, operation, and maintenance of the Service

3. Special-Care-Required Personal Information

Audio or text processed by the Service may contain special-care-required personal information such as health, medical, disability, welfare, living circumstances, or consultation content. The user understands this possibility and uses the Service with consent. When processing audio containing such information of a counterpart or third party, the user shall explain it to that person and obtain the necessary consent.

4. Provision to Third Parties Located Abroad

To the extent necessary to provide the Service, the Company may provide personal data to the following third parties located abroad.

(1) Recipients, location, purpose

OpenAI, L.L.C. (United States)
For speech recognition, translation, speech synthesis, and AI processing, audio, text, target information, and related technical information may be sent to its API. Under OpenAI’s published terms, inputs and outputs sent to business services such as the API are not used to train models unless the user or organization explicitly opts in. Under default settings, abuse-monitoring logs may be retained for up to 30 days, subject to legal and security exceptions. The Company applies safeguards appropriate to its applicable contract and configuration.

Stripe, Inc. (United States)
For credit-card payment, purchase history, payment processing, abuse prevention, and refunds/reconciliation, information needed for payment is handled. Card numbers and similar data are processed by Stripe and are not retained on the Company’s servers.

ElevenLabs, Inc. (United States)
Only when the user opts in to the voice-clone feature, the voice samples the user records, the voice characteristics (voiceprint) extracted from them, and the text used to generate translated speech are transmitted. When the user deletes the voice clone or the account, the Company deletes the voiceprint data stored with the provider.

Postmark (ActiveCampaign, LLC) (United States)
Only when Postmark is selected as the mail transport, the destination email address and message content are sent for email verification, registration completion, password reset/change completion, balance recovery, purchase/important notices, and support. The current production environment uses the Onamae.com Mail Service described below.

(2) Personal information protection systems of the relevant country

The United States does not have a comprehensive federal law equivalent to Japan’s Act on the Protection of Personal Information; personal information is regulated by sector-specific federal laws and state laws. The Company refers to information published by the Personal Information Protection Commission of Japan and the recipients’ published privacy policies, data-processing terms, and security documents, and provides users with necessary information.

(3) Protective measures taken by recipients

The Company reviews recipients’ contract terms, security documents, data-processing terms, and privacy policies, and supervises them appropriately under the Act. The Company takes contractual, technical, and operational measures so that recipients do not use the data beyond the purpose of providing the Service.

5. Outsourcing

The Company may outsource all or part of the handling of personal information for provision of the Service, payments, email delivery, server operation, maintenance, and inquiry handling, and supervises outsourcees appropriately under the Act.

The current production environment uses the Onamae.com Mail Service in Japan, provided by GMO Internet, Inc., to send confirmation, registration-completion, password-reset, password-change-completion, and related messages.

For running the application and storing its database and temporary storage area, the current production environment uses the cloud platform “Heroku” provided by Salesforce, Inc. (United States). The Company supervises this outsourcee as required by the Act. If a storage provider is added or changed, this Section will be revised and published.

6. External Transmission

The Service may transmit: (1) audio/text to OpenAI for recognition, translation, and synthesis; (1-b) recorded audio and voiceprint data to ElevenLabs when the user consents to and uses the Clone Voice feature (kept until the user deletes the clone; deletion on our side is propagated to the provider); (2) payment-related information to Stripe; (2-b) authentication information (identifier, email, and email-verification state) exchanged with Google, Apple, or Discord when registering or signing in with an external account; (3) account keys, consent state, and settings stored on the device (localStorage, etc.); and (4) logs to the Company’s servers for incident response, abuse prevention, and usage-time management. These are used for providing the Service, payments, balance management, abuse prevention, and safety. If the Company newly introduces advertising, behavioral targeting, or external analytics tags, it will publish the recipient, the information transmitted, and the purpose in this Policy or a separate schedule.

7. On-Device Storage

The Service may store account keys, consent state, language, and display settings in localStorage or similar storage. The recording feature temporarily stores unsent compressed audio in IndexedDB, capped at 64 MiB in total and 24 hours from creation, and deletes it after successful upload. Old fragments are removed when the cap is exceeded; a “process later” recording stops when it reaches the cap. Audio files the user explicitly exports are outside this automatic deletion. On a shared device, please log out or clear browser data.

8. Provision to Third Parties

Except where required by law, where the person consents, where necessary for outsourcing, or where otherwise permitted by the Act, the Company does not provide personal information to third parties.

9. Security Measures

The Company takes necessary and appropriate measures—access control, encryption, log management, permission management, and outsourcee management—to prevent leakage, loss, damage, unauthorized access, and falsification of collected information.

10. Calls, Records, and Documents

Calls. Participants’ audio is processed temporarily for translation; the audio itself is not stored. For captions and minutes, utterance text (time, display name, original text, and translation) is kept in server memory only during the call. Only when the host chooses “Keep as a record” at the end of the call is this text stored as a record linked to the host’s account (see Records below); if the host does not choose to keep it, it is deleted when the call ends and cannot be retrieved later. Before joining, a participant—even one without an account—is shown that audio is processed by AI for translation, that the host can keep the written conversation as the host’s record, and that usage time is charged to the host, and is asked for consent; audio of non-consenting participants is not processed. The consent text and its version are recorded per participant. Call connection information (participants’ display names, selected languages, and connection state) is kept only during the call and deleted when it ends; display names contained in a stored call record are handled as part of the host’s record. Natural video adjustment during a call (brightness and skin-tone adjustment) is processed only on the user’s device; no facial features or detection results are extracted, transmitted, or stored, and video is not transmitted for speech recognition or translation.

Records and documents. Transcripts, imported content, generated documents, and intermediate data are stored on the server, linked to the account, only when the user performs a record/import/save operation. Documents (summaries, reports, minutes) can be issued in the languages whose quality the Company has verified (as of this revision, 13 languages: Japanese, English, Chinese, Korean, Spanish, Portuguese, French, German, Italian, Russian, Vietnamese, Indonesian, and Hindi; up to 3 per issuance); the current list is posted on the language-selection screen and in Help. Conversation text of interpretation, venue, and call sessions is stored only when the user (the host, for calls) chooses “Keep as a record” at the end of the session; audio itself is discarded after processing and is not stored. Imported audio, video, and image files are kept in a temporary storage area on the Company’s server for transcription/text extraction and for retrying a failed job, for no longer than 24 hours after completion (7 days after failure or cancellation), and are then deleted; if the user chooses “Discard”, access is cut off and the file is deleted immediately. The temporary storage area is located on the infrastructure of the server-operation outsourcee stated in Section 5 and is encrypted at rest. When the user uses the setting that masks proper nouns in a document, the mapping between real names and placeholders is stored with that record and erased together with the record or the account; the mapping is not included in the document body or ordinary logs. Unsent recording audio is held under Section 7; an application update alone does not delete that temporary store. The user can delete stored records at any time. A deleted record is kept in the “Trash” for 30 days, during which the user can restore it or delete it permanently at once; after 30 days it is permanently deleted. When an account is deleted, the Company erases all stored records including those in the Trash (transcripts, documents, intermediate data, and proper-noun mapping tables).

11. Retention Period

The Company retains personal information for the period necessary to achieve the purpose of use, the period required by law, or a period reasonably necessary for dispute response and abuse prevention, and deletes or anonymizes unnecessary information by reasonable means.

Copies of records contained in the Company’s backups are erased no later than 35 days after the record is permanently deleted or the account is deleted. When restoring from a backup, records deleted before the restore are not restored. Ledgers of usage-time consumption, consent operations, and abuse prevention (which do not contain conversation text) are retained for the period needed for accounting and dispute response.

12. Disclosure, Correction, Suspension of Use, etc.

Users may request disclosure, correction, addition, deletion, suspension of use, erasure, or suspension of third-party provision of their own personal information held by the Company, in accordance with law. To request, contact the address below; identity verification information may be required.

13. Minors

Minors must use the Service with the consent of a parent or legal representative.

14. Revisions

The Company may revise this Policy in response to legal amendments, changes to the Service, changes to recipients, or operational needs. For material revisions, notice is given via display within the Service or other reasonable means.

15. Sign-in Assistance and Contact

The help bot on the sign-in screen displays only pre-approved fixed guidance about sign-in and registration on the user’s device. It does not send questions to generative AI or provide speculative answers. Only when the user presses “Send to Cotoa support” are the entered reply email, message, category, and necessary technical information sent to the Company’s support desk.

CoreVista-Japan, Inc.
Personal Information Protection Manager: Taketoshi Nakao
Email: support_cotoa@corevista-japan.com

Terms ・ Specified Commercial Transactions ・ Back to the Service

Revised October 25, 2026. Previous version (September 3, 2026)

Download PDF