Meeting audio and generated speech

Meeting Voice & AI Processing

This notice explains what happens to speech and meeting content when multilingual and generated meeting features are enabled, what may be stored, and the responsibilities and controls available to users and hosts.

Last updated: June 13, 2026

1. Scope

This notice applies when MeetBridge processes speech, audio, text, or a voice reference to provide multilingual meetings, generated speech, transcripts, summaries, decisions, actions, or related meeting history.

It supplements the Privacy Policy and Terms of Service. A customer agreement or workplace policy may impose additional requirements.

2. Live speech processing

When a microphone is active and speech features are enabled, audio is processed to detect spoken content and produce live captions. Caption text can be translated into participant-selected languages and can be rendered as generated audio.

Other participants may therefore see translated text or hear generated speech based on what a speaker says. Interface indicators and meeting settings should be used to understand which features are active.

3. What may be stored

Final transcripts, translations, chat, summaries, decisions, actions, participant identity, roles, and meeting timeline events may be stored in the authorized meeting history.

Audio used only for live processing is not treated as a permanent recording unless MeetBridge clearly indicates that a recording feature is enabled. Operational buffers may exist briefly to deliver the requested feature, maintain continuity, prevent abuse, or diagnose failures.

Workspace permissions determine who can access stored meeting records. Exports and downstream sharing are controlled by the customer and authorized users.

4. Generated speech

Generated speech is artificial audio produced from translated or other authorized text. It may use a default voice or, when a user has explicitly selected the option, that user's personal voice reference.

Generated speech may differ from the speaker's intent, pronunciation, emotion, timing, or wording. Participants should confirm critical statements with the original speaker.

MeetBridge may label generated audio or related content where required by law, safety standards, or product design.

5. Optional personal voice reference

Creating a personal voice reference is optional and is not required to use a default generated voice. The user records a sample that is used to create a derived voice profile for future generated speech.

A user must provide only their own voice or a voice they are expressly authorized to provide. Providing a colleague's, customer's, public figure's, or any other person's voice without authority is prohibited.

A personal voice reference must not be used for impersonation, fraud, harassment, deceptive evidence, unauthorized advertising, identity verification bypass, or misleading a listener about who actually spoke.

6. Roles and lawful basis

For customer-controlled meetings, the organization or host generally determines the purpose of processing meeting content and is responsible for an appropriate lawful basis. MeetBridge generally processes that content on the customer's instructions.

RaketaLabs may independently process limited account, security, operational, and abuse-prevention data as described in the Privacy Policy.

Where consent is required, it must be informed, specific, freely given where applicable, and capable of withdrawal for future processing. A checkbox in account onboarding does not replace meeting-specific notice or permission required by local law.

7. Host and organization responsibilities

Before enabling transcription, translation, generated speech, summarization, or recording, hosts and organizations must provide clear notice and obtain any permission required by applicable law, employment rules, contracts, or internal policy.

Hosts must not conceal active processing, invite participants under false pretenses, or use MeetBridge to conduct unlawful surveillance.

Organizations should define who can access meeting history, how long it is retained, whether exports are allowed, and which sensitive or regulated discussions are appropriate for the service.

8. Participant controls

A participant can mute their microphone, disable available media controls, leave the meeting, use a default generated voice, decline to create a personal voice reference, remove an existing optional voice reference, and exercise applicable privacy rights.

Leaving or muting stops new speech from being submitted through that participant's microphone but does not automatically erase records already created or content provided by other participants.

A request concerning customer-controlled meeting history may need to be submitted to the organization that controls the workspace.

9. Accuracy and human review

Noise, accents, overlapping speech, connectivity, terminology, and source quality can affect generated text and audio. Outputs may be incomplete, delayed, or incorrect.

Do not rely on a transcript, translation, summary, action item, or generated voice as the sole source for legal, medical, financial, employment, immigration, safety, or similarly high-impact decisions.

For critical communication, use qualified human review or a certified professional appropriate to the context.

10. Sensitive and regulated content

Participants should avoid disclosing sensitive personal data, trade secrets, credentials, payment data, health details, or regulated information unless the organization has approved the use, established a lawful basis, and applied appropriate safeguards.

MeetBridge is not represented as a dedicated emergency, medical-device, legal-interpretation, financial-advice, or regulated decision-making service.

11. Service providers and international handling

MeetBridge may use vetted service providers to perform hosting, storage, security, speech processing, translation, generated speech, communications, and support functions. Providers receive only the data needed for the requested function and are subject to contractual controls.

Processing may occur outside the participant's country. Where required, international transfers use recognized legal safeguards.

12. Retention and deletion

Live operational audio is kept no longer than needed for the active feature unless recording is clearly enabled. Stored meeting records follow customer settings, plan capabilities, legal obligations, and deletion or backup cycles.

Optional voice references and derived profiles are retained until removed, the account is deleted, the feature is disabled, or a shorter contractual period applies.

Removing a voice reference stops future use after operational deletion completes. It does not retract generated speech already delivered or automatically delete authorized meeting records.

13. Contact and updates

Questions about meeting processing, voice references, participant rights, or organization controls can be submitted through the MeetBridge contact page.

We may update this notice for legal, safety, operational, or product changes. Material changes will be communicated appropriately and the effective date will be updated.