Logos Talk — Privacy Policy & Terms of Service

This Privacy Policy explains how Logos collects, processes, and protects your data when using our debate analytics platform.

1. Core Commitment & Optional Recording

While built on the principle of data minimization—measuring speaking time and keyword occurrences via ephemeral processing by default—our platform gives users the option to retain debate session recordings securely:

• Ephemeral Processing by Default: Real-time speaker diarization and custom keyword detection normally operate strictly on transient audio streams in volatile memory (RAM), discarding raw audio immediately after analysis.
• Optional Recording: Users may explicitly choose to record and archive their debate sessions. When enabled by the user, recordings are stored securely within their private workspace.
• Strict Confidentiality: No one else will have access to your saved debate recordings. They remain strictly private and accessible only to authorized account holders.

2. Information We Collect & Manage

• Voice Training Samples (Speaker Identification):
  • Collection: Organizers capture brief voice samples (5 to 10 seconds per speaker) solely to construct localized mathematical voiceprints for distinguishing participants.
  • Control: Voice samples and derived acoustic vectors are stored under the client’s private workspace and can be deleted at any time via the dashboard.
• Custom Keyword Configurations: Client-specified lists of trigger words, technical terms, or phrases selected for real-time occurrence counts.
• Debate Analytics & Metadata:
  • Collection: Cumulative speaking durations, turn frequencies, timestamps, participant labels, session titles, dates, frequency counts of selected keywords, and optional debate audio recordings (if saving is enabled by the user).
  • Control: All session statistics and recordings remain attached exclusively to the organizer’s account. Clients retain total ownership and can delete session metadata or recordings at any time.
• Account Credentials & Invoicing: Basic administrative details (name, email, password hash) and transaction receipts handled securely by third-party payment gateways.

3. Purpose of Processing

Your data is processed strictly to:

• Train event-specific speaker profiles to attribute speaking time accurately.
• Match real-time phonemes against client-designated keywords to calculate occurrence metrics.
• Record and store debate audio only when explicitly chosen and enabled by the user.
• Provide downloadable analytics dashboards, optional debate playback, and aggregate reporting to the organizer.
• Authenticate authorized account holders and maintain service integrity.

We do not sell or broker data, nor do we use proprietary debate keyword lists, optional recordings, or participant audio samples to train generalized public AI models.

4. Application Performance and Technical Limitations

The performance of the diarization function (speaker identification) provided by the application depends on hardware and environmental variables. The accuracy of the identification is limited by the following factors:

• Sound quality: the overall acoustic environment, background noise, reverberation, and the distance between speakers.
• Internal hardware and native software processing: The recording depends on the physical quality of the phone's built-in microphone, as well as the operating system's audio processing algorithms.
• Number of participants: The complexity of the algorithmic processing increases with the number of speakers present. Voice overlap (crosstalk) can temporarily reduce the identification accuracy.
• Optimization via external hardware: For optimal results, the use of an external tabletop microphone (such as a ReSpeak USB Mic Array connected to the device) is strongly recommended. With this type of equipment, the application can efficiently distinguish between up to ten simultaneous speakers.

5. Storage, Security & Retention

• Security: Data in transit is secured with TLS 1.3 encryption, and account metadata/recordings at rest are secured with AES-256 encryption. Access to stored debate recordings is strictly restricted to the account owner.
• Retention Schedule:
  • Training Voice Samples: Stored until manually cleared by the organizer or upon account termination.
  • Debate Analytics & Optional Recordings: Kept within your account for ongoing historical review and playback until manually purged by you.
  • Account Deletion: Closing your account triggers the permanent deletion of all stored voiceprints, keyword lists, session analytics, and recordings within 30 days.

6. User Rights & Data Control

Under applicable global data protection regulations (including GDPR, CCPA/CPRA, and biometric/voice privacy frameworks), clients and participants have the right to:

• Export Data: Download comprehensive CSV or PDF summaries of speaking times and keyword occurrences (along with stored audio files, if recording was enabled).
• Immediate Purge: Delete specific voice training files, custom keyword lists, debate records, or saved recordings directly from the interface.
• Erasure & Inquiries: Request full deletion of account holdings by contacting privacy@logos-4u.com

7. Contact Us

For questions regarding this policy or technical data safeguards, contact:
Email: legal@logos-4u.com
Address: [Company Legal Address]

---

Terms of Service: Client Obligations Regarding Participant Consent & Notice

1. Scope of Client Responsibility

When using our services to monitor debates, panel sessions, or meetings, the Client acts as the independent data controller (or business) with respect to the personal data and biometric identifiers of all event participants, speakers, and attendees. The Client acknowledges and agrees that the Client possesses direct privity and sole operational control over communication with participants.

2. Mandatory Notice and Consent Requirements

Prior to utilizing the Application during any session, the Client represents, warrants, and covenants that it shall:

• Advance Disclosure: Clearly and conspicuously notify all participants prior to the event that the Application will be utilized to measure speaking durations, track occurrences of pre-designated keywords, and—if applicable—record the debate session.
• 5-Second Voice Calibration: Inform participants before capturing any voice sample that the 5-second recording is collected solely to create an acoustic profile for speaker identification, will not be shared externally, and may be deleted at any time upon request.
• Obtain Valid Consent: Procure all legally necessary prior written, verbal, or electronic consents from each participant required under applicable data protection, voice privacy, and wiretapping laws (including, but not limited to, the EU/UK General Data Protection Regulation (GDPR), the California Consumer Privacy Act/CPRA, the Illinois Biometric Information Privacy Act (BIPA), and applicable two-party or all-party audio recording consent statutes).
• Provide an Opt-Out Mechanism: Make reasonable accommodations for any participant who declines voice calibration, keyword attribution, or audio recording (e.g., categorizing their speaking time under an uncalibrated/anonymous profile or excluding them from recorded audio streams).

3. Processing & Recording Representations

The Client shall accurately represent the functionality of the Application to participants. Specifically, the Client shall confirm that:

1. Live speech is analyzed ephemerally in volatile memory, but may be recorded and stored privately only if the user explicitly chooses to save the debate;
2. When audio recordings are saved by the user, no one else will have access to them;
3. Only derivative metadata (speaking time allocations and aggregate keyword tallies), the 5-second calibration samples, and user-elected recordings are retained in the Client’s account; and
4. Voice training samples, session analytics, and stored recordings can be purged by the Client at any time.

4. Deletion and Subject Rights Requests

The Client is solely responsible for fulfilling data subject requests made by its participants. If a participant requests the erasure of their voice calibration sample, stored audio, or the anonymization of their metadata, the Client shall promptly execute the deletion via the dashboard or submit a verified written request to our support team.

5. Indemnification

The Client agrees to defend, indemnify, and hold harmless the Company, its affiliates, directors, officers, and employees against any third-party claims, regulatory enforcement actions, penalties, damages, or legal expenses (including reasonable attorneys' fees) arising out of or resulting from the Client’s failure to provide adequate notice, obtain valid participant consents, or comply with applicable biometric privacy or recording consent laws.