Privacy Policy

In short: what happens to your data.

• We collect the minimum: your email and name from Google (or your Telegram account), your content in the studio, the history of your credit grants and charges, and — at sign-in — your IP address and browser (for security); if you book a demo or write to us from the site — what you enter in the form; if you came through one of our ad links — which one; where you came to us from (site, page, link tags).
• No third-party analytics, ad pixels or ad networks. Our own ad links (vinistudio.app/c/…) remember for 30 days which link you came through, and we count visits to them ourselves, as totals — see §2. Only strictly necessary cookies — which is why there is no cookie banner.
• Your prompts and media go to AI providers only to fulfil your request.
• We do not train models on your content and do not sell personal data.
• The server is in the EU (Germany). Account and data deletion — by emailing [email protected].

1. Who is responsible for your data?

The party responsible for your personal data (its "controller", in GDPR terms) is the operator of the Vini Studio service (vinistudio.app). The service is in closed beta; the legal entity details will be published in this section before paid plans launch. For any data question: [email protected].

This policy is part of the Terms of Service.

2. What data do we collect?

Account. When you sign in with Google — your email, name, and Google account identifier. With Telegram — your Telegram identifier and username. We do not store passwords — authentication is handled by Google and Telegram.

Optional YouTube connection. Vini Studio uses YouTube API Services. If you connect a channel, we receive its identifier, name, handle and, when available, profile image, together with the permissions you grant and OAuth access and refresh tokens. We use youtube.readonly to identify and display the connected channel and check upload processing, and youtube.upload to upload the videos you choose. We process the video, title, description and publication settings you submit, and store the publication result, including the YouTube video identifier and status. Connecting YouTube is separate from signing in with Google.

Sessions. Signing in creates a session valid for 7 days; the IP address and browser string (user-agent) are stored with it — this is needed for account security.

Your content. Uploads (photos, videos, audio), prompts, generation outputs, and the projects themselves (the block diagrams you assemble in the studio) are stored on our server so the studio works and you don't lose your work.

Credits. The history of grants and charges: amount, generation type, model. We do not collect bank card data — when payments launch, they will be processed by a payment provider; your card details are never stored with us.

Interface language — so we don't have to ask every time.

Account profile and contact permission. If you complete your profile, we store your role, goals and description of your idea, plus your company or project and how you heard about the studio if you provide them. The owner sees these answers alongside your account and the campaign or invitation through which you registered. Your contact for personal communication is stored separately, together with your permission, its date and the version of the consent text. Signing in with Telegram or Google does not itself give permission to contact you. Contact details remain optional even when a profile is required to enter the studio. The owner may write to you about your idea and experience with the studio; partners do not receive your profile or contact details. You can edit your answers and withdraw permission in your account: withdrawal ends permission to contact you, its date is stored, and you keep any bonus already credited. Bonus records support balance accounting and prevent duplicate grants.

Demo requests and messages from the site. If you book a demo on the "Demo" page or write to us through the "Contact" page, we receive what you enter in the form: your name, your contact in the channel you choose (Telegram, WhatsApp, email, or LinkedIn), who you are and what you want to make, your message, and — if you provide them — your company, how you heard about us, and an invite code. If you arrived through a partner's invite link, that link is attached to the request. If you send a request while signed in, it is linked to your account. The page language and the date of this policy's version are stored with it. We use this data only to reply about your selected topic: content production, team setup, a demo or private access to future subscriptions; we send no newsletters and write to you only about your request.

Ad links. If you open one of our ad links (an address like vinistudio.app/c/…), your browser remembers it for 30 days with the avg_campaign cookie. If you create an account within that time, we save with it which link you came through (the channel and the campaign) and add the bonus that link promised. For visits to the link itself we store only an hourly total, without any link to you as a person. At the moment of a visit we look at the IP address, the browser string and the link cookie — to leave out bots, repeats and flooding — and store none of them.

Where you came from. When you open our site, we record where you came to us from: the site you came from (the address without parameters), the page you landed on, the utm tags of ad links and the link's ?ref= tag, the name of the ad platform's click tag (for example, gclid), the device type, the system, the browser and the app the link was opened in, the language and the country according to Cloudflare. We do not store any other address parameters, the IP address, the browser string or the values of ad platforms' click tags. Before you create an account, this record sits in the avg_origin cookie in your browser and is not linked to you; when you create an account, it is saved with the account, and the cookie is deleted when you sign in.

Service-side access. The service administrator can see your account data (email or username, credit balance, and charge history) and, for support and moderation, may open your studio in read-only mode — without being able to change or run anything on your behalf.

Subscription interest. We count daily views of the subscription offer and request-link clicks as aggregate totals, separately from saved requests. These counters contain no visitor identifiers, IP addresses or browser strings and do not represent purchases. The IP address is used temporarily in server memory to limit request frequency.

3. Why, and on what legal basis?

Purpose	Legal basis (GDPR)
Operating the service: account, generations, project storage	Contract (Art. 6(1)(b))
Security: sessions, IP, abuse prevention	Legitimate interest (Art. 6(1)(f))
Replying to your request for a demo or your message from the site	Steps taken at your request before entering into a contract (Art. 6(1)(b)) and our legitimate interest in answering you (Art. 6(1)(f))
Credit accounting and charges	Contract (Art. 6(1)(b))
Understanding which advertising works: aggregate visit counts for our ad links and the link an account was created through	Legitimate interest (Art. 6(1)(f))
Understanding where people come from: which site, page and ad they arrived through	Legitimate interest (Art. 6(1)(f))
Ad link bonus	Contract (Art. 6(1)(b))
Voice cloning	Consent (Art. 6(1)(a)): for your own voice — your explicit upload of the audio; for another person's voice — that person's prior explicit consent, which you must obtain before uploading (Terms, section 4)
Retaining specific records required by law (e.g. financial accounting)	Legal obligation (Art. 6(1)(c))

4. What goes to AI providers?

Generation is performed by third-party AI model and compute providers (image, video, audio, and text models; speech transcription). We pass them your prompts and media only to fulfil your specific request — this is technically necessary; without it, generation is impossible.

• Faces. Uploaded photos containing faces are passed to providers to produce the frame or clip you requested; a provider may temporarily retain the submitted materials and outputs under its own data-processing policy. We do not create or store biometric identification templates (faceprints, voiceprints), and the service may not be used for biometric identification of people.
• Voice. Audio for voice cloning is passed to a speech-synthesis provider only upon your explicit action.
• Providers may process data outside the EU — see section 7.

5. Do we train models? Do we sell data?

No and no. We do not train AI models on your content (we have no models of our own), and we do not sell or share personal data for third-party advertising or marketing.

YouTube channel information and OAuth tokens are used to support channel connection and the publication features you request. OAuth access and refresh tokens are not included in requests to AI providers. We do not use data obtained through Google or YouTube APIs to train AI models. When you publish, your selected video and publication details are sent to YouTube. Google's processing is described in the Google Privacy Policy. Our use and transfer of Google API data follow the Google API Services User Data Policy, including its Limited Use requirements.

6. What cookies do we use?

Only strictly necessary ones — which is why there is no cookie banner.

Cookie	Purpose	Lifetime
avg_sid	Sign-in session	7 days
avg_oauth	Protects Google sign-in and social account connections, including YouTube, against request forgery (CSRF); set only during authorization	10 minutes
avg_invite	Partner invite: remembers the invite link you arrived through until you create an account	30 days
avg_campaign	Ad link: remembers which of our ad links you arrived through until you create an account; deleted when you sign in	30 days
avg_origin	Where you came from: remembers which site and ad brought you to us until you create an account; deleted when you sign in	90 days
avg_recipe	Recipe hand-off: remembers the recipe you opened while you sign in, so that exactly that recipe opens afterwards	30 minutes
lang, langBanner	Interface language	1 year

There are no third-party analytics, ad pixels, or third-party trackers on the site. To operate the service we keep internal usage records (generations, charges, studio activity) and aggregate visit counts for our ad links — they are never shared. External resources used by the pages: the Telegram sign-in widget (on sign-in pages and in account settings when linking Telegram) and YouTube video previews in the blog and articles (the preview image loads from YouTube's servers; the youtube-nocookie player itself loads only after a click).

7. Where is the data stored?

Our server is located in the EU (Germany, a Hetzner data centre). AI model providers may process the requests passed to them outside the EU/EEA — such transfers rely on the mechanisms provided by the GDPR (the EU Standard Contractual Clauses, SCC, or equivalent safeguards) in line with each provider's data-processing terms. We will send you the current list of providers and a copy of the applicable safeguards on request at [email protected].

Notifications to us via Telegram. When a new account signs up, we receive a notification in Telegram: the account's name, email or username (or, if there is none, the account ID), the sign-in method, and — if you arrived through an invite — the name of the partner who invited you, or, if you arrived through one of our ad links, the name of that ad campaign and, if a partner placed the link, the partner's name. When an account requests a top-up, we receive its name and account ID. This data is stored on Telegram's servers, which may be located outside the EU. The legal basis is our legitimate interest (Art. 6(1)(f) GDPR): to know promptly that someone new has joined and to answer a top-up request.

Requests from the site. When you book a demo or send a message from the site, we receive a notification in Telegram with everything you entered in the form and — if you arrived through an invite — the name of the partner who invited you and how the invite is tied to the request (by code, by link, or through your account). This data is stored on Telegram's servers, which may be located outside the EU.

7.1. How we protect your data

We use the following safeguards to protect personal data, including sensitive Google and YouTube user data, against unauthorized access and disclosure:

• Encryption in transit. Connections to our public website and our server's connections to Google and YouTube APIs use HTTPS/TLS.
• Encrypted YouTube credentials. YouTube OAuth access and refresh tokens are encrypted with AES-256-GCM before they are stored in our database. The encryption key is held in server configuration, separately from the database records. These tokens are used on the server and are excluded from the account data returned to your browser.
• Access controls. Access to connected-channel records and publication actions requires authentication and is checked against account ownership. Service-side support and moderation access is described in section 2.
• Protected sessions and account linking. On the production service, session and OAuth cookies use Secure, HttpOnly and SameSite protections. Google and YouTube authorization uses short-lived, single-use state bound to the initiating browser; connecting a YouTube channel also requires the same signed-in account when authorization returns.

8. How long do we keep data?

• Account data and content — while the account is active.
• Sessions — 7 days.
• Demo requests and messages from the site — up to 12 months after our last exchange if you do not become a user of the service; to have them deleted sooner, email [email protected].
• Deletion. Email [email protected] — we will delete your account and associated data within 30 days; backups are purged shortly after. We may retain specific records longer where the law requires it (e.g. financial accounting).

YouTube access. You can disconnect a channel in your account's social accounts settings. This deletes the selected connection record and its stored tokens from our active connections and requests revocation of the authorization at Google. You can also revoke access directly in your Google Account permissions. Disconnecting does not delete videos already uploaded to YouTube or your studio projects and publication history. To request deletion of data retained by Vini Studio, contact [email protected]; to manage videos on YouTube, use YouTube's own controls.

9. Your rights

Under the GDPR you can: access your data and get a copy, rectify it, erase it, restrict processing, object to processing based on legitimate interest, and withdraw consent (this does not affect the lawfulness of processing before withdrawal). A single email to [email protected] is enough for any request.

If you believe we are violating your rights, you can lodge a complaint with the data protection supervisory authority in your country of residence.

10. Children, changes, contact

The service is intended for people aged 18 or older; we do not knowingly collect children's data.

We may update this policy — we will announce significant changes by email or an in-service notice. The date of the current version is shown at the top of the page.

Data questions: [email protected].