# Porus AI Privacy Policy Effective 24 October 2026. Version 2026-10-24-v2. ## Who is responsible and how to contact us Porus AI provides the service. Send privacy, consent-withdrawal, rights, or grievance requests to Porus AI at legal@porus.ai. Include your account email and, where relevant, the job reference so we can identify and address your request. ## What we collect and why - Account data: name where supplied, verified contact email, a phone number required for email/password signup, Cognito account identifier, and account-verification status, to create and secure the account. Password handling is through Amazon Cognito; the Studio application does not store your plaintext password. Google signup does not require a phone number. - Job data: uploaded videos, audio and voices within them, filenames, glossary terms, language selections, job status, and eventual outputs, to carry out and support the requested dubbing job. Material may contain other people's data; only submit it with appropriate permissions. - Consent evidence: authenticated account identifier, policy and notice versions, document hashes, the individual declarations you accepted, server receipt time, and request identifiers. Where collected for evidence, source IP and user-agent are disclosed and retained under the reviewed schedule. A client-supplied time is only an assertion and is not treated as independently verified. - Operational data: necessary authentication, error, abuse, and security events, to operate the service, investigate faults, and respond to misuse. - Communications: support and rights requests and job-related correspondence. The account and job processing described above is separate from optional future training, marketing, or speaker-verification purposes. A purpose change requires an updated notice and permission where required; it is not authorized simply by accepting these account terms. ## Human access and service providers The current service includes manual processing. Authorized Porus personnel may access a submitted video and its metadata to perform your job. Files are not made public or placed in a public download directory. Hosting, identity, storage, and job records currently use AWS services. The live AWS configuration uses the Hyderabad region, ap-south-2, for Cognito, S3 media, Lambda, and DynamoDB. Our website uses CloudFront. Do not interpret this as a promise that all support, network, or service-provider processing is exclusively within India. Google account login is available through a separate Amazon Cognito user pool. When you choose it, Google shares the identity details requested on its sign-in screen, including an account identifier, email and verification status, and profile name where available. Your Google password stays with Google. We request basic identity scopes only, not access to Gmail, Drive or contacts. Google's own privacy terms govern its authentication service. Existing email/password accounts stay in their original Cognito pool. A Google login creates a separate profile; matching email addresses do not automatically merge accounts or transfer uploaded videos. Use your original login method for its earlier jobs. Social identities and job ownership are scoped to the provider pool and account identifier. A contact email must be verified before uploads. Microsoft account login is not enabled in this release. Additional providers or processing purposes will be disclosed before introduction, with further permission where required. ## Voice data and training Recordings may identify a speaker and can require particular care. We do not currently offer a verified voice-cloning enrollment flow, create reusable voice profiles under this account consent, or use customer material for model training. Do not submit a speaker's voice for cloning on the assumption that your own account checkbox grants that person's permission. Any future voice enrollment needs its own notice about what is created, purpose, retention, sharing, verification, and withdrawal. Training would require separate optional permission and a distinct data-handling workflow. No permission for voice cloning, advertising, sale of voice profiles, or model training is inferred from accepting the Terms or submitting a job. ## Storage, security, and retention Uploads are private and encrypted at rest in S3. Authentication gates the upload API; job access is scoped to the account. Time-limited upload and download links are used. Only authorized operators should access media for manual processing. These controls reduce risk but are not a guarantee against every security event. Media retention in the current alpha is managed manually. Our service-copy targets are 30 days from upload for source videos, 30 days from completion for delivered results, and 7 days after completion, cancellation, or failure for temporary voice material and intermediate files. Queued jobs may require a documented extension. These are operating targets, not automatic expiry settings or a guarantee that every copy disappears at the same time. Automated media expiry, voice-data destruction, and self-service deletion are not enabled. Request deletion at legal@porus.ai. We assess the request, identify the relevant storage and manual-processing copies, and explain any preservation exception. We may retain restricted evidence or security records where necessary for a justified purpose or required by applicable law. We do not use these exceptions as permission to train on your media. Consent evidence is kept separately from media. It records the account, accepted versions, declarations, and receipt time. Retention is limited to the period needed to demonstrate authorization, handle disputes, and meet legal duties. Backups and legally preserved copies may outlast deletion from active storage; we will explain relevant limitations when handling a request. ## Your choices and requests Write to legal@porus.ai from your account email with the relevant job reference to request access, correction, export, deletion, cancellation, or withdrawal of a specified permission. We may ask for proportionate evidence to verify a request and protect other people. We will explain any applicable retention exception or reason a request cannot be fulfilled. Response deadlines must follow applicable law. We will acknowledge and assess your request and explain the next steps. Self-service export and deletion are not currently implemented. Withdrawal can stop future processing of the relevant purpose. It may make a pending dubbing job impossible to complete. It does not retract content already published elsewhere by you or others. Essential security and legally required records may need to be retained under a restricted purpose. Regulatory complaint routes depend on applicable law and the commencement of the relevant provisions. ## Children, browser storage, and communications The service is adults-only, 18+. We do not offer a verified parental-consent mechanism. Contact legal@porus.ai if a child has submitted information. We will investigate and take appropriate action. The website uses browser storage needed for sign-in and unfinished registration. There is no marketing/training opt-in bundled into account acceptance. A cookie and storage inventory must be completed before adding analytics, advertising, or non-essential tracking. Transactional job messages are separate from marketing. ## Changes and regional review We will publish versions and explain significant changes. New processing purposes will not silently expand an earlier permission. Applicable statutory rights continue to apply. If the service expands to new purposes or markets, we will provide additional notices or permissions where required. This policy describes the current alpha and is not a certification of compliance with every privacy or biometric law.