SaydiVoice Privacy Policy: how your data is handled

Why data protection matters for an AI voice service

With most online tools the data you submit is just text. With a text-to-speech service it also includes voice recordings — which can identify a specific person and count as personal data under the law. So the question is not only whether a service is secure, but:

• Where is my recording and text stored, and for how long?
• Is any of it used to train AI?
• Who inside the system can access it?
• Can I delete it, and what do I lose if I do?

The SaydiVoice privacy policy answers those four questions. Below is a summary of the points that matter most.

What data does SaydiVoice collect?

We group data into four categories:

• Account information: email, display name and profile picture if you sign in with Google, plus your plan status and remaining allowance.
• Data you submit: text, document files, subtitles, voice recordings, audio or video you upload for dubbing.
• Usage and technical data: generation history, characters used per window, the voices and languages you pick, IP address, device type and error codes.
• Transaction data: plan purchased, amount, time and payment status — but never your card number.

We collect only what is needed to run the service, we do not buy data from third parties and we do not collect anything beyond the categories above.

What is the data used for?

SaydiVoice processes data to provide the service you request, manage accounts and allowances, take payment and meet accounting and tax duties, support customers, prevent abuse and protect the infrastructure, and improve quality using aggregated or anonymised usage data.

We do not sell your personal data and we do not use it for advertising outside the service.

Are my voice samples used to train AI?

No. This is our most important commitment and the question we are asked most often:

• Voice samples you upload for cloning are not used to train Saydi AI's models.
• The audio you generate is not used to train models either.
• Samples are used only to create the voice you asked for and stay linked to your account until you delete that voice or your account.
• Speech generation runs on our own infrastructure; your data is not sent to a third-party AI model for training.

For features that require translation or language processing, data may be sent to an AI provider under a data-protection commitment, solely to return the result of your request.

Who can access my data?

We apply a need-to-know rule: operational staff only access the data a specific task requires, and every access goes through role-based controls.

A small number of partners process data to the minimum extent necessary:

• Infrastructure and storage: servers, databases and the GPU infrastructure that generates speech.
• Authentication: the sign-in and session service.
• Payments: the transaction processor; we never store card numbers.
• Analytics and support: usage measurement tools and the chat channel when you contact us.

Every processor acting on our behalf commits to confidentiality and may act only on our instructions.

What rights do I have over my personal data?

Under Decree 13/2023/ND-CP on personal data protection you have the right to be informed, to consent or refuse, to withdraw consent, to access, correct and delete your data, to restrict or object to processing, to receive your data for portability, to complain and to claim compensation for damage.

Many of these you can exercise yourself in the studio — delete a cloned voice, delete your generation history, update your account details. For anything else, contact us through our contact page and we respond within 72 working hours.

Some rights may be limited where fulfilling them would infringe another person's rights or where the law requires us to retain records, such as accounting and tax files.

How does SaydiVoice protect data?

The main measures in place:

• Encryption in transit: all website and API connections use HTTPS/WSS.
• Expiring session tokens, role-based access control and a need-to-know rule for operational staff.
• Per-account and per-IP rate limiting plus anomaly detection.
• Secrets and configuration kept out of application source and out of browser bundles.
• Regular backups and system error monitoring.

No system is absolutely secure. If you spot a security risk, tell us immediately so we can act early.

How long is my data kept?

We keep data for as long as it is needed for the purposes described:

• Account data: while the account is active and for a reasonable period afterwards to handle disputes.
• Voice samples and cloned voices: until you delete them.
• Generation and dubbing history: according to your settings, deletable at any time.
• Transaction data: for the periods required by accounting, tax and anti-fraud law.
• Technical data and logs: briefly, for operations and security.

When a retention period ends, the data is deleted or anonymised.

Cookies and analytics

SaydiVoice uses cookies sparingly: essential cookies to keep you signed in and remember your language, and analytics cookies to see which pages are used and which errors occur — with some of our analytics designed to run without personal tracking cookies.

You can block or delete cookies in your browser settings. Note that blocking essential cookies may prevent you from signing in.

Quick summary of the points that matter

• We collect only what is needed to run the service and we do not sell data.
• Voice samples and generated audio are not used to train AI models.
• You control your input data; Saydi AI processes it on your instructions.
• Clone a voice only if you own it or hold the data subject's consent.
• You can access, correct and delete your data, withdraw consent and complain.
• Many rights can be exercised in the studio without waiting for support.
• We respond to personal-data requests within 72 working hours.

The full, legally binding text is the SaydiVoice Privacy Policy. Where this article and the policy differ, the policy prevails.