Privacy policy

The controller under the General Data Protection Regulation (GDPR) is:

The operating company is a wholly owned subsidiary of the holding company ILG Group GmbH. This structure is part of our transparency: our stance is open (“We are not neutral. We are on your side.”), and our AI-generated content is labelled as AI-generated.

This policy applies to the website commonvoice.media, including every to.cv link that forwards to it.

• No account, no profile. Our chat works without signing in. We don't build user profiles and don't collect names or contact details unless you enter them in a form yourself.

  Section 5

• No advertising, no third-party trackers, no selling of data. Usage statistics exist only if you agree to them, and they stay with us.

  Section 4

• Open about AI. The chat is an AI conversation, and our videos are AI-generated. We say so openly.

  Section 5

• Honest about what is stored. This policy describes what our systems do today, not only what we plan.

  Section 10

When our videos appear on external platforms (for example TikTok, Instagram, YouTube, Mastodon), watching, commenting and sharing there are governed by that platform's privacy policy. We receive no personal data about individual viewers from the platforms, at most aggregated statistics. We don't publish on any platform yet; this list will be added when we do.

The link names the video, never a person. A short code such as to.cv/xxxxx stands for a video. Everyone who watches the same video uses the same link; it contains no user identifier.

Hosting and server logs

The website, the chat and our servers are hosted by IONOS SE, Elgendorfer Str. 57, 56410 Montabaur, Germany (processing on our behalf under Art. 28 GDPR). Data-centre location: to be added. When you visit, our servers process technically necessary data in logs: IP address, time, the page or interface requested, browser/user agent, device and operating system, referrer and language setting. The legal basis is Art. 6(1)(f) GDPR (secure, reliable operation). How long the logs are kept is in section 10. Technical error alerts to our internal team may contain IP addresses and requested paths.

What is stored in your browser

• A cookie with your language choice (NEXT_LOCALE), so the website keeps your language.
• In local storage (kept until you clear it in your browser): your choice of light or dark design, your privacy setting and the app's display settings (collapsed sidebar, video player height, the split between video and chat).
• Your story draft, if you write one in the app: title, text and topic, in local storage so nothing is lost if the tab closes. Your name and email address are never stored in it. The draft is deleted as soon as you send the story or clear the fields. On a shared device, please clear the fields before you leave.
• In session storage (deleted as soon as you close the tab): during a chat, the conversation and a session identifier; your likes; whether you closed the notice about our app; and, for staff, the sign-in to the internal Studio.

This storage is necessary for the features you ask for (§ 25(2) no. 2 TDDDG, the German telecommunications and digital services data protection act). It doesn't leave your browser, except the chat session identifier, which is sent with each chat message. We don't set advertising or tracking cookies.

Optional usage statistics (only with your consent)

If you agree, the website sends to our own server which videos are watched, how far, whether they are finished, shared or opened in the chat. If a chat session exists, its identifier may be included. We use this to understand which stories work; it goes to no third party. The legal basis is your consent (Art. 6(1)(a) GDPR, § 25(1) TDDDG). You can withdraw it at any time via “Privacy settings” at the bottom of every page. Retention: see section 10.

Likes: when you like a video, we count it so the public number is right. No session identifier is sent with it.

How the chat works

The chat is an openly declared AI conversation with the character from the video. There are no accounts and no sign-in. A chat session belongs to your browser tab: within that session, the chat knows which videos you have watched and what you have talked about. When you close the tab, the next session starts from zero.

Processing during the conversation

To generate replies, we pass your messages and the conversation so far to these providers:

• Mistral AI SAS, Paris, France: generating the replies;
• Tabularis GmbH, Heilbronn, Germany: assessing the mood and emotion of a message.

We run the conversation control (n8n) ourselves, on our own servers. The legal basis is Art. 6(1)(f) GDPR (providing the chat). We exclude any use of your content to train the models; contractual confirmation: to be added.

We don't ask for names, contact details or other personal information in the chat. Please don't enter any of your own accord either.

What is stored

• On our servers: session information (session identifier, videos watched) and the last ten messages of a session (yours and the chat's) in our database. For the ongoing conversation, the system also keeps the context in a cache for up to 24 hours and the last 20 turns in memory. Retention in the database: see section 10.
• In your browser: the conversation for display, until you close the tab.

What doesn't exist: no user profiles, no political categorisation, no linking with other records, no sharing for advertising.

What we analyse conversations for

• An editorial learning loop: which topics, questions and moods come up, so we can improve our content.
• Quality checks of the replies (samples against verified fact sheets).
• Detecting misuse, after the fact and never in real time against individual sessions.

Not for: advertising, sale, sharing or political profiling.

Notes

• Young people: our service is meant for people aged 13 and over. We recommend that parents accompany their children's media use.
• Acute crises: the chat is not therapy and not an emergency service. If there are signs of an acute emergency, it points to established help services (for example Nummer gegen Kummer, TelefonSeelsorge).

You can send us a story through the “Share your story” form. We process: the text of your story, optionally a title and a topic, your choice to stay anonymous or be credited by name (and that name, if you give one), your consent, and optionally an email address.

• Use: our team reviews the story. To publish it, we turn it into a script and the script into an AI-generated video. For that, the text is sent to Mistral AI (script writing); for multi-part productions, OpenRouter, Inc. (USA) may be used. Runway AI, Inc. (USA) and Higgsfield Inc. (USA) generate the video from the script; finished videos are stored with Cloudinary (USA) and possibly on PeerTube (see section 9).
• Email address: optional; we use it only to ask follow-up questions about your submission.
• Legal basis: your consent (Art. 6(1)(a) GDPR), which you can withdraw at any time.
• Retention: see section 10.

Please don't include names or details by which you or others could be recognised.

When you give us feedback or report a chat reply, we process: the rating or category, the video or link concerned, your free text and, if you give it, your email address. The purpose is to improve and check our content; the legal basis is Art. 6(1)(f) GDPR. Retention: see section 10.

If you email us (at website@commonvoice.media), we process your details only to handle your request (Art. 6(1)(f) or (b) GDPR, depending on the matter). Once it is dealt with and any retention obligations have expired, we delete the correspondence (see section 10).

We have, or are concluding, data processing agreements with our service providers (Art. 28 GDPR). Transfers to countries outside the EU are based on to be added. We do not sell data. We disclose data to authorities only where legally required.

You have the right of access (Art. 15), rectification (Art. 16), erasure (Art. 17), restriction of processing (Art. 18), data portability (Art. 20) and objection (Art. 21 GDPR). You can withdraw any consent you have given (for example to usage statistics or for a submission) at any time, with effect for the future.

Because we don't keep accounts, we can't link chat sessions to a person. If you give us your session identifier or details, we check what we can find and delete it.

You also have the right to complain to a data protection supervisory authority. The authority responsible for us is the Berlin Commissioner for Data Protection and Freedom of Information.

Send requests to: website@commonvoice.media

We protect our systems to the state of the art, including TLS encryption and minimal access rights. The chat's behaviour lives in inspectable, versioned texts; we don't train models on your conversations.

We update this policy when our processing changes, for example when a new service is added. The current version is always at commonvoice.media/en/datenschutz. The German version at commonvoice.media/de/datenschutz is the binding one.