Privacy policy

Oruk’s own inference services discard request audio after processing. Resonance 2 WebSocket sessions buffer audio in memory until reset or session closure; an in-flight analysis keeps its committed audio until it finishes. Raw audio is not saved in the durable Resonance 2 replay record. Optional speaker diarization uses task processors that temporarily retain uploaded audio for up to 48 hours. Completed Resonance 2 result payloads, including expression scores, are stored for durable settlement and authorized same-ID replay. Replay is available for 24 hours; this is not a physical-deletion deadline. These replay records contain no raw audio or transcript and are separate from request metadata. The speaker diarization processor retains its job output for up to 24 hours.

Oruk does not use customer audio or outputs to train, fine-tune, or evaluate models without your explicit written agreement.

Request identifiers, model, task, audio duration, usage, and timestamps are retained for billing, security, and support. These records do not contain your audio or transcript.

Enabling optional speaker diarization sends audio to a task processor. File requests use its Media API: uploaded audio is deleted within 48 hours and speaker-label job output within 24 hours under the provider’s published retention policy. Live sessions stream audio to its WebSocket service. These processing paths are not a promise of zero retention across all providers.

Standard speech inference is operated in the United States. Optional task processing may occur outside the United States. If your application requires a specific processing region, retention term, or deployment boundary, confirm the complete configuration with Oruk before sending audio.

Feature-specific processing and deletion details are available during a security review. Optional or custom processing terms must be confirmed for the configuration you use. See security and deployment.

API audio

Oruk’s own inference services discard request audio after processing. Resonance 2 WebSocket sessions buffer audio in memory until reset or session closure; an in-flight analysis keeps its committed audio until it finishes. Raw audio is not saved in the durable Resonance 2 replay record. Optional speaker diarization uses task processors that temporarily retain uploaded audio for up to 48 hours.

Demo and playground file uploads

Demo and playground files selected for batch testing are staged in a private temporary cache before you click Run. Upload handles expire after 10 minutes. Replacing a file or closing the demo requests deletion sooner. Staged files are not used for training.

Transcripts and acoustic annotations

Completed Resonance 2 result payloads, including expression scores, are stored for durable settlement and authorized same-ID replay. Replay is available for 24 hours; this is not a physical-deletion deadline. These replay records contain no raw audio or transcript and are separate from request metadata. The speaker diarization processor retains its job output for up to 24 hours.

Usage and request metadata

Retained for billing, abuse prevention, audit, and support. These records do not contain the recording or its transcript.

Account details

Retained while your account is active. Access, correction, and deletion requests are handled by our team, subject to required record retention.

Passwords and API keys

Stored in non-recoverable form. A complete API key is shown once.

Consented attribution identifiers

Eligible for measurement for up to 90 days after the most recent consented capture; cleared on withdrawal or by scheduled cleanup after expiry.

Subscription and financial records

Retained as required for accounting, tax compliance, and dispute handling.

Usage and financial records do not have a universal automatic deletion period. Our team handles requests in light of the account, transaction, and legal retention requirements.

When you create an account, we collect your name and email address, your authentication details, and any optional company, role, or use-case information you provide. We keep subscription, payment, invoice, and usage records to operate your account and calculate charges.

For new signups, we record the terms version you accepted, the acceptance time, and whether you signed up with email or Google. These records document your agreement and are separate from optional marketing preferences.

Card details, billing addresses, and tax information are collected through secure hosted checkout. Oruk stores customer, payment, and invoice identifiers but does not receive or store complete card numbers.

Operational logs may include IP address, browser user agent, request path, timestamp, and status for security and reliability. API recordings and model outputs are excluded from marketing measurement.

Batch audio testing uses a short-lived, essential cookie to keep temporary uploads private to your browser session. It expires after 10 minutes and is not used for advertising or analytics.

Advertising and analytics storage default to denied until your cookie preference permits them. Choosing Essential only keeps them denied. Server-side advertising reports also require permission and are suppressed when the browser sends Global Privacy Control.

With consent, first-party cookies can retain advertising click identifiers, campaign parameters, a landing-page address, and a random visitor identifier for up to 90 days. We may store consented analytics client and session identifiers to link a public-site visit to account registration, the first successful API request, and a subscription or payment event.

Advertising reports can contain the event time, a random event identifier, IP address, browser user agent, a public page address without query parameters, available advertising identifiers, and a conversion amount where applicable. Some advertising reports require a corresponding ad-click identifier. Consented homepage measurement can also include visits without an ad click.

For ChatGPT ads, permitted server-side reports to OpenAI contain an OpenAI ad-click identifier, a random event identifier, the event time, our public site address, and account registration, trial, first successful API request, or paid subscription milestones. Paid subscription reports include the amount actually paid. These reports omit email addresses and hashes, IP addresses, browser user agents, and customer content, and opt out of future user-level personalization.

We do not include your name, email address, audio, transcripts, emotion labels, filenames, API keys, or account content in these reports. Measurement runs on public acquisition pages, with operational account and billing records handled separately.

Use Cookie preferences in the footer to withdraw consent. Withdrawal clears linked first-party identifiers and stops queued exports. Expired identifiers are cleared by scheduled cleanup after their eligibility period. Information already reported to an advertising or analytics service follows that service's retention settings. You may also request deletion of attribution associated with your account.

We do not sell personal data or customer content. Data is disclosed only as necessary to deliver the selected service, operate accounts and billing, perform measurement you have permitted, protect the service, or comply with law. Contracted processing is limited to the relevant service purpose. Customer audio and outputs are not disclosed for advertising or model training without your explicit written agreement.

If you accept the organizational customer-reference permission, we may also display your organization’s name and logo to identify it as a customer. This does not permit publishing personal account details or customer content. You can opt out by emailing access@oruk.ai.

You can request access, correction, or deletion of your personal data by emailing access@oruk.ai. We may verify your identity before acting. Required accounting, legal, and security records may be retained. Requests can cover retained Resonance 2 results as well as account, operational, and attribution records. The authorized replay window does not define a physical-deletion deadline.

Traffic is encrypted in transit, account records are encrypted at rest, and production services use authenticated access controls. See the security page for current controls, limitations, and incident reporting.

We update this page and its effective date when this policy changes. Material changes to API content-retention or model-training commitments will be communicated to account holders before taking effect.