Cotoa Talk & Keep — Terms of Service

Effective: June 5, 2026 / Last revised: October 25, 2026 (v6: storage of call transcripts at the host’s choice, temporary storage of imported files, the on-device temporary-storage cap, deletion and the Trash, what happens when the balance runs out, pause/resume consumption, and how supported languages are published. Previous revision: October 17, 2026) / Provider: CoreVista-Japan, Inc.

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

1. Application

These Terms of Service (the “Terms”) set out the conditions of use of the voice interpretation and records service “Cotoa Talk & Keep” (the “Service”) provided by CoreVista-Japan, Inc. (the “Company”). Users use the Service only after agreeing to these Terms, the Privacy Policy, the Notation based on the Act on Specified Commercial Transactions, and any notices shown within the Service.

2. The Service

The Service performs speech recognition, translation, caption display, and speech synthesis based on the user’s language settings, using AI. Because the Service involves automated AI processing, the Company does not guarantee the completeness, accuracy, immediacy, or fitness for a particular purpose of the translation results, recognition results, captions, or audio output.

3. Nature of AI Translation and Cautions

The user understands that translation results may contain errors, delays, omissions, recognition mistakes, inability to capture audio, and issues arising from the communication or device environment. In medical, welfare, nursing, legal, administrative, contractual, emergency, or other situations with serious impact on life, body, or property, do not rely solely on translation results; obtain confirmation by a professional, qualified person, or human as necessary.

4. Management of Account/Seat Keys

The Service is used via account keys, seat keys, URLs, or QR codes issued to purchasers or administrators. The user is responsible for managing such keys and must not share, transfer, lend, publish, or leak them. Operations performed using such keys may be deemed to be performed by the manager or user, except where the Company is willfully or grossly negligent.

Email registration is complete when the user finishes confirmation on the page linked from the Company’s confirmation email. Email-and-password sign-in is unavailable before confirmation. The Google and Discord buttons are shared sign-up/sign-in entry points: first use creates an account, and later use signs in with the same external account.

When the user requests a password reset, the Company sends a time-limited, single-use reset link to the registered email. After a successful change, the Company sends a completion notice to the same address and rotates the existing account key.

5. Audio Processing and Third-Party Consent

The user agrees that audio, text, and target information are processed for translation, speech recognition, speech synthesis, billing, balance management, and abuse prevention. When processing the audio of a counterpart, co-present person, or other third party, the user shall explain the use of the Service to that person and obtain the necessary consent. The user understands that audio may contain special-care-required personal information such as health, medical, disability, welfare, or living-circumstance information, and shall use the Service only with the consent of the relevant person.

6. Fees, Purchases, and Balance

The Service is available with free time (Article 6-2) and one-time paid balance top-ups. The standard tax-inclusive prices are 5 minutes ¥280, 15 minutes ¥680, 30 minutes ¥1,180, 1 hour ¥1,980, 3 hours ¥5,400, 6 hours ¥9,900, and 12 hours ¥18,000, with no automatic renewal. The short packs (5, 15, and 30 minutes) can also be added from the interpretation screen. Purchased balance does not expire. One balance minute equals one actual minute of simultaneous interpretation, conversation, earphone, or venue use with standard voice. Captions only consume 0.50x. Remote calls keep the other participant’s original voice audible and show the translation as captions only, without synthesized or cloned readout; they consume 0.50x per minute of audio processed (30 seconds of balance per processed minute). User-selected cloned voice is available for simultaneous interpretation at 1.50x. Recording/transcription consumes 0.05x of recorded time. Confirmed historical charges are not changed. Document generation is quoted before execution.

Realtime interpretation, captions, and cloned voice have a 60-actual-second minimum at start and then accumulate by connection time in seconds. If the user pauses and then resumes the same interpretation, the minimum unit is not charged again (resuming in a different interpretation mode is treated as a new start and is shown before it begins). Processed-audio multipliers round a fraction of a balance second up; recording/transcription applies 0.05x to the cumulative recording duration and drops fractions below one balance second. A document estimate is a cap and a failed generation is refunded. The Company keeps the confirmation and consumption record. Before each feature starts, the Service shows the remaining balance and the cap for that use (5 minutes per session for an account holding only Free Time). If the balance runs out during use, the Company stops processing new audio and finalizes what was already accepted: in a remote call, only the translated captions stop while the original-voice call and camera video continue; in recording, on-device recording continues while only transcription stops. The Company does not switch the user to another mode or another account without the user’s approval because the balance ran out. Accounts with positive purchased balance at the August 15, 2026 migration receive a one-time addition equal to 50% of that purchased balance; free allowance is excluded. Purchased time cannot be exchanged for cash or transferred, resold, or lent. Refunds are limited to cases required by law or stated in the commercial-transaction notice.

6-2. Free Time

The Company grants free usage time (“Free Time”) to users who create an account by registering and verifying an email address, with Google, or with another external account whose email is verified, such as Discord. No free allowance is provided through instantly issued seat keys. (1) During the beta campaign (through December 1, 2026, 23:59:59 JST), the first 15 accounts each day (in order of completed verification) receive 5 minutes of Free Time every month; slots are released daily at 7:00, 12:00, and 20:00 JST, 5 slots each. Unused Free Time carries over, up to a maximum holding of 10 minutes, and an account holding only Free Time may run a single interpretation session for up to 5 minutes. (2) Accounts created during the beta campaign after the daily limit is reached receive a one-time 5 minutes on creation (verification) with no monthly grant. (2-b) Accounts created before the August 24, 2026 revision that have completed email verification also receive the monthly 5 minutes during the beta campaign, within the daily slots. (3) Accounts created on or after December 2, 2026 receive a one-time 2 minutes on creation (verification). (4) Free Time is consumed before purchased balance; multipliers, minimum units, and rounding follow Article 6. Free Time cannot be exchanged for cash or transferred and is excluded from the migration protection in Article 6. (5) The Company may revoke Free Time in case of fraudulent multiple accounts or other violations of these Terms. (6) The conditions, amount, daily limit, and campaign period of Free Time may be changed or ended without notice; changes apply to Free Time granted after the change.

7. Prohibited Acts

The user must not: (1) violate laws or public order; (2) infringe the rights, privacy, reputation, or credit of third parties; (3) process a third party’s audio without consent; (4) create a cloned voice using any voice other than the user’s own (celebrities, third parties, existing recordings, etc.), or use a cloned voice to impersonate another person, defraud, deceive, create deepfakes, or otherwise mislead; (5) obtain, record, or use conversations in a manner that violates applicable law (including unlawful interception or misuse of secret recording); (6) falsify transcripts, summaries, minutes, or reports generated by the Service and present them to third parties as authentic records; (7) wrongfully obtain, share, sell, transfer, or publish account/seat keys, URLs, or QR codes; (8) circumvent or interfere with billing, balance management, usage limits, or security; (9) place excessive load on the Service, servers, API, or network; (10) perform reverse engineering, scraping, unauthorized access, bot use, vulnerability probing, or similar acts; (11) cause damage to the Company or third parties; or (12) other acts the Company deems inappropriate.

7-2. Clone Voice

The Service offers an optional feature that extracts voice characteristics (a voiceprint) from audio the user records and creates and stores a synthetic voice used to render translated speech (“Clone Voice”). Use of this feature requires the user’s separate consent to this Article. (1) The user may register only the user’s own voice, and must not register the voice of anyone else (celebrities, acquaintances, third parties, existing recordings, etc.). (2) The user must not use a Clone Voice for impersonation, fraud, deception, deepfakes, defamation, or any other unlawful or improper purpose. (3) The user represents and warrants that the registered audio and the generated Clone Voice do not infringe any third-party rights (including moral rights, publicity rights, portrait rights, intellectual-property rights, and privacy rights). (4) Rights in the audio and the Clone Voice belong to the user; the Company handles them only to the extent necessary to provide the feature (creation, storage, and generation of translated speech) and does not use them for any other purpose (including AI model training). The feature relies on an external provider (see Article 5 and the Privacy Policy). (5) Up to 2 Clone Voices may be stored per account. The user may delete them at any time; upon deletion, the Company erases them without delay, including the voiceprint data stored with the external provider. (6) If a violation of this Article is suspected, the Company may suspend or delete Clone Voices, suspend the Service, or disable the account without prior notice. (7) If the user’s violation of this Article causes damage, claims, or costs (including attorneys’ fees) to the Company or third parties, the user shall defend, indemnify, and hold the Company harmless.

7-3. Calls and Minutes

The Service may offer a call feature in which multiple participants receive interpretation at the same time. (1) Processing time consumed during a call is charged to the account of the user who opened the call (the “host”); invited participants can join without an account. (2) A participant’s audio is translated only if the participant consents to audio processing before joining; audio of non-consenting participants is not processed. (3) The host shall explain to participants in advance that audio is processed for translation and that the host can keep the written remarks as a record, and shall obtain any required consent; invited participants are shown the same information on the screen before joining. The user is responsible for complying with recording-related laws. (4) During a call, remarks (time, display name, original text, and translation) are held in server memory only during the call, for captions and minutes. Only when the host chooses “Keep as a record” at the end of the call are they stored as Record Data linked to the host’s account (Article 7-4). If the host does not choose to keep them, or the call ends without a choice, the remarks are deleted when the call ends and cannot be retrieved later; the Company does not store generated minutes without the user’s instruction. (4-b) A stored call record can be viewed, edited, and deleted only by the host; a participant who wishes to have their own remarks deleted may ask the host or the Company’s contact (Article 14). (5) Minutes are AI-generated and may contain errors, omissions, or speaker mix-ups; the Company does not guarantee their accuracy, and users must verify important matters themselves. (6) Calls may fail to connect depending on network conditions; the Company does not guarantee call connectivity or quality.

7-4. Records and Documents

The Service can, at the user’s operation, store transcripts of recordings, conversation text of interpretation, venue, and call sessions, the content of imported files (audio, video, text, images), and AI-generated summaries, reports, and minutes (“Record Data”). Record Data is stored only when the user performs a record, import, or save operation; conversation text of interpretation, venue, and call sessions is stored only when the user (the host, for calls) chooses to keep it at the end of the session, and is otherwise not stored. Audio itself is discarded after processing and is not stored, except that audio, video, and image files the user imports are kept in a temporary storage area on the Company’s server, solely for transcription or 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; if the user chooses “Discard”, the Company cuts off access and deletes the file immediately. Files in the temporary storage area are not the user’s stored property and cannot be retrieved after that period; keep the original files yourself. Unsent compressed audio may be held temporarily on the user’s device to survive transmission failure or interruption, capped at 64 MiB and 24 hours, and deleted after successful upload; when the cap is reached, recording stops or the oldest unsent fragments are removed; an explicitly exported file is outside this temporary store. The user is responsible for third-party consent and applicable recording law. AI-generated documents may contain errors, omissions, and mix-ups of speakers, figures, or decisions, and have no legal effect; verify important use against the source. The user may delete Record Data at any time. Deleted Record Data is kept in the “Trash” for 30 days so that it can be restored after a mistaken deletion; during that period the user may restore it or delete it permanently at once, and after 30 days it is permanently deleted. Copies contained in the Company’s backups are erased no later than 35 days after permanent deletion. When an account is deleted, the Company erases stored Record Data including items in the Trash, and backup copies within the same period. Documents (summaries, reports, minutes) can be issued only in 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—in up to 3 languages per issuance; documents cannot be issued in other languages. The current list of supported languages is posted on the language-selection screen and in Help, where additions and changes are announced; a language already supported will not be withdrawn without prior notice. Time consumption follows Article 6.

Recording, transmission, or saving may stop when a device locks or the app moves into the background, microphone permission is revoked, the network disconnects, storage is insufficient, or browser storage is cleared. Unsaved content may not be recoverable. Make a short test recording before important use, and check both the save-completed indication and the saved content afterwards. Retain original audio/video files separately and export important records and documents for safekeeping. Temporary storage and resume features do not guarantee recovery of every unsent audio segment. This paragraph does not exempt the Company from its responsibility under Article 9.

7-5. Withdrawal and Account Deletion

The user may stop using the Service at any time. An account with a verified email address can be deleted from “Me → Account → Data handling → Delete account” in the app by opening the confirmation link sent to the registered email (valid for 1 hour). Deletion permanently erases the remaining time (both free and purchased), Record Data (including items in the Trash), and Clone Voices (including voiceprint data stored with the external provider), and cannot be undone. Backup copies are erased within the period stated in Article 7-4. Unused purchased time is not refunded except as provided in the notation based on the Act on Specified Commercial Transactions. Accounts without a registered email address can request deletion via the in-app contact form.

8. Change, Suspension, and Termination

The Company may change, suspend, halt, or terminate all or part of the Service when necessary for maintenance, incident response, security, legal compliance, changes to external services, or other reasons. Except in unavoidable emergencies, the Company will endeavor to give prior notice of changes, suspensions, halts, or terminations that materially affect users, via display within the Service or other reasonable means.

9. Disclaimer and Limitation of Liability

The Company does not warrant that the Service always operates normally, that translations are always accurate, that it is usable on all devices/browsers/networks, or that it fits the user’s particular purpose. Where the Company is liable for damage to the user due to default, tort, or causes attributable to the Company, except in cases of the Company’s willful misconduct or gross negligence, the Company’s liability is limited to ordinary and direct damages actually incurred. Only where the Company is slightly negligent, the total liability is capped at the total Service fees the user paid to the Company in the six months before the date the damage occurred; for users with no payment in that period, the cap is JPY 1,500. This cap does not apply in cases of the Company’s willful misconduct or gross negligence.

10. Changes to the Terms

The Company may change these Terms when (1) the change conforms to the general interest of users, or (2) the change does not contradict the purpose of the Terms and is reasonable in light of the necessity, the appropriateness of the revised content, and other circumstances. The Company will notify the revised content and effective date via display within the Service or other reasonable means; for material changes, notice is given before the effective date. Version 6 (effective October 25, 2026) is announced by posting it within the Service and takes effect on the day it is posted; at the first use after it takes effect, the Service asks the user to confirm agreement to the revised Terms.

11. Use by Minors

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

12. Exclusion of Antisocial Forces

The user represents and warrants that the user is not an antisocial force and will not allow antisocial forces to use the Service.

13. Governing Law and Jurisdiction

These Terms are governed by Japanese law. Any dispute between the Company and the user shall be subject to the Osaka District Court as the court of first instance by agreement; provided that, where the user is a consumer under the Consumer Contract Act, jurisdiction permitted by law is not precluded.

14. Contact

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

Privacy Policy ・ Specified Commercial Transactions ・ Back to the Service

Revised to v6 on October 25, 2026. Previous version (v5) · Versions before October 19, 2026

Download PDF