Privacy Policy

Privacy that matches how OnadaMeet actually works.

This policy explains the distinction between end-to-end encrypted live media and optional AI processing, as well as access, retention, deletion, human review and participant-notice safeguards.

01

Information we process

OnadaMeet processes account and profile information, workspace membership and settings, meeting and classroom schedules, participant and attendance records, chat and operational metadata, support and security information, and billing-related service records needed to operate the service.

When an authorized customer uses AI Meeting Intelligence, OnadaMeet may also process meeting notes, transcripts or other authorized meeting content and generated records such as summaries, discussion highlights, speaker contributions, decisions, action items, unresolved issues, risks and conclusions.

02

End-to-end encryption and AI processing

When OnadaMeet identifies a meeting as end-to-end encrypted, the live meeting media is encrypted so intermediary services do not receive the media encryption keys required to read that media. E2EE protects the live-media path; it does not prevent a participant or authorized feature from intentionally exporting, recording, transcribing, copying or submitting content to another feature or recipient.

AI Notes is a separate, optional processing path. To generate an AI-assisted meeting record, authorized meeting content such as a transcript or notes must be deliberately made available for that processing. Enabling AI Notes does not give an AI provider the ability to decrypt the E2EE media stream.

03

AI Meeting Intelligence, notice and human review

For manual AI Notes, an authenticated meeting host or authorized administrator must affirm that they are authorized to submit the meeting content and are responsible for providing any notice or obtaining any permission or consent required by applicable law, contract or organizational policy.

Automated live AI processing is designed to require participant-facing notice and a persistent in-meeting indicator while processing is active. Participant choices are recorded as the factual events they represent; an OnadaMeet acknowledgement record is not a representation that every form of consent required in every jurisdiction has been obtained.

AI-generated records can be incomplete or inaccurate. They require human review and are not authoritative legal records merely because OnadaMeet generated or stored them. Users should not rely on unreviewed AI output for legal, employment, financial, academic, disciplinary, compliance, safety or other consequential decisions.

04

Speaker attribution and biometrics

OnadaMeet is designed to use authenticated participant/session identity or explicit transcript metadata for speaker attribution when reliable information is available. If reliable attribution is absent, the service should preserve uncertainty rather than guess a person's identity.

AI Meeting Intelligence is not intended to create reusable biometric voiceprints merely to identify meeting speakers, and OnadaMeet does not instruct the AI system to infer identity from voice characteristics, accent, demographic traits or contextual guesswork.

05

Service providers and subprocessors

OnadaMeet may use contracted infrastructure, database, authentication, payment, email, media, security, analytics and AI-processing providers to operate requested features. When a customer invokes an AI feature, the content necessary to provide that feature may be processed by the configured AI provider subject to applicable agreements and provider settings.

We do not describe customer meeting content as inaccessible to a processor when a customer has deliberately invoked a feature that requires that processor to handle the submitted content.

06

Access, retention and deletion

Meeting transcripts and AI-generated records are subject to workspace and meeting access controls. AI-record mutation controls are restricted to authorized meeting hosts or workspace administrators; ordinary workspace membership alone does not authorize modification or deletion of another host's AI record.

Authorized hosts or administrators can delete a stored transcript or the full AI meeting record through supported product controls. Transcript retention can be left under manual control or configured for supported automatic deletion periods. Deletion events may retain limited audit metadata such as meeting, actor, scope, reason and time without retaining the deleted transcript content.

Retention may also be affected by account status, backups, security needs, contractual commitments, legal requirements and dispute-resolution obligations. We avoid promising a deletion period that the product does not technically enforce.

07

Security and data minimization

OnadaMeet uses technical and organizational safeguards designed to protect account, meeting and workspace information. AI prompts should include only content needed for the requested meeting-intelligence task and should not intentionally include unrelated authentication secrets, payment information, private encryption keys or unrelated private records.

Meeting transcripts are treated as untrusted model input. Provider failures are not replaced with fabricated successful summaries, and routine error handling should not intentionally log full meeting transcripts.

08

Your choices and privacy rights

You may update supported account information, manage workspace access, configure supported transcript retention, delete supported AI records, export available records or contact us about a privacy request. Depending on your location, you may have rights to access, correct, delete, restrict or object to certain processing.

For privacy questions or requests, contact privacy@onadameet.com. OnadaMeet is operated by Onada Innovative LLC.

09

Changes to this policy

If OnadaMeet materially changes how existing meeting content is used, we will provide appropriate notice and seek consent where required rather than relying on a materially inconsistent hidden interpretation of an earlier privacy promise.

Effective and last updated: August 22, 2026.