Privacy Policy

1. Controller

Droidtech e.K., owner Michael Rauen, Auf der Eichelsbach 19, 54533 Hasborn, Germany. Phone +49 6574 227 0937, e-mail mail@klickdemo.com. No data protection officer has been appointed because the legal requirements (Section 38 BDSG, Art. 37 GDPR) are not met. The owner is the contact for data protection matters.

This policy covers the website klickdemo.com, the Klickdemo application (app), the embed player and the interfaces. The legal bases are the GDPR, the German Federal Data Protection Act (BDSG) and the German Telecommunications Digital Services Data Protection Act (TDDDG). Klickdemo is offered to businesses only; personal data arises mainly for the people who use an account on behalf of a company, for website visitors and for viewers of embedded videos.

Two roles. For your account data, your payments and the operation of the platform we are the controller. For the content of your projects (credentials of a demo account, recordings, screenshots, Excel files, uploaded documents and other knowledge sources, texts) and for the view counts of the embed player on your website we process data on your behalf as a processor (Art. 28 GDPR). For this you conclude a data processing agreement with us in the application. This policy nevertheless describes the technical flows in full so that you know what happens to the data.

2. Website

Hosting and server logs. The website and the application run on a server of Host Europe GmbH, Hansestraße 111, 51149 Cologne, in a data centre in Germany (processor, Art. 28 GDPR). On each request the server processes technically necessary data: IP address, time, requested address, browser identifier, referrer, status code. The purposes are delivery, security and troubleshooting (Art. 6(1)(f) GDPR). Server logs are deleted after 14 days at the latest.

Consent banner. On your first visit the website asks whether, beyond the technically necessary processing, you allow services in the categories “Statistics”, “Marketing” and “External media”. Declining is as easy as accepting and has no disadvantages; until you choose, everything stays off. The website stores your choice in your browser (klickdemo.consent, localStorage, 12 months) and asks again afterwards. Via “Cookie settings” in the footer of every page you can change or withdraw your choice at any time; withdrawal takes effect for the future (Art. 7(3) GDPR). The legal basis for the services requiring consent is Art. 6(1)(a) GDPR and Section 25(1) TDDDG.

Appearance and language hint. If you choose the “Light” or “Dark” appearance in the website header or in the application, your browser stores this choice (klickdemo.theme, localStorage, until you choose “Match system” or clear the site data). If you close the hint about the English version on the website, your browser remembers that (klickdemo.langhint) so the hint does not show again. Both are strictly necessary for the function you requested, never leave your browser and need no consent (Section 25(2) no. 2 TDDDG).

Google Tag Manager and Consent Mode v2. Only after you allow “Statistics” or “Marketing” does the website load Google Tag Manager (Google Ireland Limited, Gordon House, Barrow Street, Dublin 4, Ireland). Before your choice no Google script is loaded; the Consent Mode v2 signals (analytics_storage, ad_storage, ad_user_data, ad_personalization) default to “denied” and are set to “granted” only with your consent. Tag Manager itself sets no cookies and stores no data about you; it loads the services you have allowed.

Statistics: Google Analytics 4. With your consent to “Statistics”, Google Analytics 4 measures which pages are viewed for how long, with which device and browser, from which country and via which route you arrived, as well as scrolling, clicks on external links, file downloads and interaction with embedded videos. Google Analytics sets cookies (_ga, _ga_*, lifetime up to 13 months) and processes the data on our behalf (Google Ads Data Processing Terms). According to Google, Google Analytics does not log or store the IP addresses of users in the EU; your IP address is used only to derive your approximate location (country, region, city). Google Signals and sharing with other Google services (including for modelling and industry benchmarks) are switched off; event data retention at Google is set to 14 months. Data may be transferred to Google LLC in the USA; the basis is the adequacy decision for the EU-US Data Privacy Framework, under which Google LLC is certified, and the EU standard contractual clauses (section 13). If you withdraw your consent to “Statistics”, the website deletes the _ga and _ga_* cookies of our domain in your browser.

Marketing: Google Ads. We currently run no ads; Tag Manager contains no Google Ads tag, and consent to “Marketing” only loads Tag Manager (above) without Google Ads collecting any data. We will use Google Ads only once we run ads, and only with your consent to “Marketing”; we will update this policy then. The intended setup is: with your consent to “Marketing”, Google Ads measures whether a click on one of our ads led to a registration (conversion measurement) and may show you Klickdemo ads on Google services and partner sites (remarketing). For this Google sets cookies (such as _gcl_au, lifetime 90 days) and processes IP address, browser identifier and the pages visited. The controllers are Google Ireland Limited and, for the transfer to the USA, Google LLC (Data Privacy Framework, standard contractual clauses). Without your consent there is no conversion measurement and no remarketing.

External media. Klickdemo videos embedded on the website run through our own player without cookies (section 8) and need no consent. YouTube videos are loaded only after you allow “External media” or load a video individually; Google LLC then receives your IP address and sets its own cookies. We use the domain youtube-nocookie.com. Google’s privacy policy: https://policies.google.com/privacy.

Referral link (?ref=). If you arrive through a partner’s referral link, the website appends that partner’s code and the time of the visit to the registration links so that the referral can be attributed when you sign up. Without your consent this happens only in the memory of the open page: the referral then counts only if you register in the same visit, and nothing is stored in your browser. If you have consented to the “Statistics” category, the website additionally stores the code and the time of the first visit in your browser (klickdemo.ref, klickdemo.refAt, localStorage, 90 days); a later visit through a different referral link does not overwrite the entry, the first contact counts. The legal basis for the storage is your consent (Section 25(1) TDDDG, Art. 6(1)(a) GDPR); if you withdraw it, the website deletes the entry. The code belongs to the partner, not to you. We set no cookie for this; as long as you do not register, nothing is transmitted to us. The attribution at registration serves the settlement of the partner commission (Art. 6(1)(f) GDPR).

Fonts and scripts. Fonts and scripts of the website are served from our own server; no connection to Google Fonts or other font services is made. The only exception is Cloudflare Turnstile, which the teaser page loads only after you have selected a file (below).

Free teaser and bot protection. Without an account you can have a short teaser video generated from an Excel file on the website; a teaser from a website address is only available with a free account. For the teaser without an account you provide the file, your company name and your business e-mail address (Art. 6(1)(b) GDPR); we are the controller for these data. To protect against automated requests the page checks Cloudflare Turnstile (Cloudflare, Inc., 101 Townsend Street, San Francisco, CA 94107, USA, with Cloudflare Germany GmbH as its European contact). Turnstile is loaded only after you have selected a file; it processes IP address, browser characteristics and interaction patterns, sets no tracking cookie and builds no profile (Art. 6(1)(f) GDPR, Section 25(2) no. 2 TDDDG). To limit the teaser to one per IP address and day and one per company domain and day (for addresses of general e-mail providers, one per address and day), we store the IP address as a salted hash for 24 hours. Disposable e-mail addresses receive no teaser. You confirm your address with a six-digit code that we send you by e-mail. Right after the upload our server analyses the file without AI (number of sheets, formulas and cells); nothing is sent to an AI model before you have confirmed your address. After that, the structure, formulas and texts of the file go to the model (section 5): the software first replaces numbers, dates and boolean values by their kind (such as “number” or “currency”); text cells such as labels and headings are transmitted, cut to 60 characters and, for long text columns, limited to a few entries per column. We send you the link to the finished video by e-mail. We delete the uploaded file, the knowledge base created from it, the recording and the video 2 days (48 hours) after the video was made available or, if no video is made, 2 days after the upload; at the same time we delete the company name, e-mail address and file name. We keep the domain of your e-mail address (the part after the @) for as long as we offer the teaser, to enforce the limit per company domain and to evaluate the use of the teaser (Art. 6(1)(f) GDPR). If you create an account with the same e-mail address within the period, the project becomes part of your account and the rules for accounts apply (sections 4 and 14).

Contact. If you write to us by e-mail or through the contact form, we process name, e-mail address, company and message to handle the request (Art. 6(1)(b) or (f) GDPR). The form checks a hidden field and the time between display and sending; there is no tracking. We delete requests sent through the contact form six months after they were settled and those recognised as spam after 30 days; if a request leads to a quote, the quote’s retention obligation applies (section 14).

3. Account, sign-in, team

Registration. On registration we process business e-mail address, display name, company name, your confirmation that you act as a business, time and IP address of registration and the chosen language. The purposes are providing the account and performing the contract (Art. 6(1)(b) GDPR). We store the registration IP address to prevent multiple accounts and misuse of the free allowance (Art. 6(1)(f) GDPR). If the account is created through a referral link, we store the partner code and the time to settle the partner commission; the partner receives no personal data, only the number and revenue of referred accounts without names.

Origin of the registration. To measure how customers find us, website and application pass on where you came from until the registration: campaign details from the link (utm_source, utm_medium, utm_campaign, utm_content, utm_term), the click identifier of a Google ad (gclid, gbraid, wbraid), a partner code, the landing page without further address parts and only the name of the referring website (such as www.google.de), not its full address. On the way from the website into the application these details are part of the link; the application keeps them in memory only. Without your consent the website only uses the details of the page currently open and stores nothing in your browser. If you have consented to the “Statistics” category, the website remembers for at most 90 days where you came from on your first and on your latest visit (klickdemo.attr, localStorage); the legal basis for this is your consent (Section 25(1) TDDDG, Art. 6(1)(a) GDPR), and the website deletes the entry if you withdraw it. The details reach us only with the registration (or with the upload of a file for the free teaser; if you then create an account with the same address they move into the account, otherwise we delete them with the other teaser details); we store them with the account for as long as it exists and evaluate them only aggregated per channel (Art. 6(1)(f) GDPR). We set no cookie for this, store no IP address for it and build no device profile.

Passwordless sign-in. Klickdemo has no passwords. To sign in we send a short-lived one-time code to your e-mail address; the code is valid for a few minutes and for one use only, and failed attempts are limited. After sign-in we store a session token (refresh token) as a hash with time of first and last use and the browser identifier so that you can see and end your sessions in the account. Sessions end when you sign out or after prolonged inactivity.

Team. The account owner can invite further people by e-mail and give them the roles Owner, Admin, Editor or Reviewer. Invitees see the account name, the projects and the names and e-mail addresses of the other members. For member data (name, e-mail address, role, activity in the account) the legal basis is performance of the contract with the company (Art. 6(1)(b) GDPR) and our legitimate interest in traceable collaboration (Art. 6(1)(f) GDPR). Who approved a script, generated a video or changed settings, and when, is logged in the account; we keep this log for twelve months.

Billing details. For paid services we record company, billing address, country, VAT identification number where applicable, purchase order reference and billing e-mail address (Art. 6(1)(b) and (c) GDPR). We validate the VAT identification number through the payment provider against the EU VAT Information Exchange System (VIES).

4. Projects: credentials, recordings, files, knowledge sources

A project describes the software, website or file that Klickdemo is to explain. For the following data you are the controller and we are the processor; the data processing agreement governs the details.

Demo-account credentials. For a web application with a login you store credentials of a demo account (username, password, TOTP secret or API token where applicable). We store them exclusively encrypted (AES-256-GCM, key separate from the database, unique random nonce per record). They are decrypted only for the duration of a job into the memory of the recording worker, which types them into the login fields. The AI model never receives credentials: the agent works with placeholders that the software substitutes only when typing. Credentials are deleted as soon as you delete the source or the project or close your account. We expressly recommend a test account without real customer or employee data; the application points this out when you create a source. Free accounts (teaser) do not accept credentials.

Project mail address. Each project receives its own e-mail address with us, which you can enter in the demo account of the software to be recorded so that the agent can receive sign-in codes and invitation links. Messages to this address are received by Cloudflare Email Routing and forwarded immediately to our server, which stores them (section 7). The agent receives only the extracted code or link from a message, never the whole message. We delete received messages 30 days after receipt, at the latest with the project.

Exploration, screenshots, knowledge base. The agent operates your software in an automated browser, reads the page structure (accessibility tree), takes screenshots and builds a knowledge base from them: pages, elements, flows, terms and facts with their evidence. Before page texts and page structure go to the model, the software replaces credentials and typical sensitive patterns (e-mail addresses, IBANs, card numbers) with placeholders; in screenshots for the model it blacks out the places where these details appear. In addition you can define your own redaction rules for your account and for each project (texts, patterns, CSS selectors, elements of the knowledge base, and presets for prices, e-mail addresses and names in typical name fields). Whatever such a rule matches, the software replaces in page texts, blacks out in screenshots before they go to the model, replaces in the texts of Excel files and additional knowledge sources, and makes unrecognisable in the video recording. Whatever no rule and no pattern matches, the model can see if it is visible in the demo account. You define which actions the agent may perform; the default “show only” allows no creating or deleting. What the agent creates is registered and, where possible, removed again after the run. Knowledge-base screenshots remain until a new exploration replaces them or you delete the project.

Recordings and videos. For each video the worker records the browser and produces the final file with synthetic voice, captions and chapters. For every version of a video we store its raw take (silent picture track, narration track and action log) so that you can later change voice, pronunciation, captions or cards without a new recording; it is deleted together with the version. Other intermediate files and the audit screenshots of individual actions are deleted no later than 30 days after the run. Of every video we keep the current and the previous version per language; older versions are deleted 30 days after two newer ones replaced them. Otherwise finished videos, captions, transcripts, posters and metadata remain until you delete them or the project; after the end of a contract, for free accounts and for unsettled storage credits the periods in section 14 apply. To measure and bill storage we store the location, type, size and storage class of every file, a daily total per account, and per video the time of the last view or download (without reference to the person who viewed; Art. 6(1)(b) GDPR).

Excel files. Spreadsheet files (xlsx, xlsm, xls, ods, csv) are analysed without AI (sheets, formulas, references, formats). Structure, formulas and texts of the file go to the model: the software first replaces numbers, dates and boolean values by their kind; text cells such as labels are transmitted, shortened and, for long text columns, limited to a few entries per column. The model never receives the file itself. The file remains until you delete it or the project.

Additional knowledge sources. Besides the main source you can give a project further sources so that the agent understands your product better:

• Public websites (such as a product page, help centre, release notes): the software fetches pages of the same domain up to the page limit of your plan and takes a few screenshots. The text of the pages goes to the model for summarising. The pages serve only as knowledge; they are not recorded. Addresses in private networks are blocked.
• Documents (PDF, Word, PowerPoint, Markdown, text): we store the file in our object storage (section 7). The software extracts the text without AI and shows you the page count and an estimate; only after your confirmation do text sections go to the model, never the file itself and no images from it. Scanned documents without text are not processed (no text recognition).
• Notes: short texts to the agent that enter the knowledge base and later jobs without a model call of their own.

The software hands content from websites and documents to the model as untrusted data, not as instructions. We delete a source and its entries in the knowledge base when you delete the source or the project. Please upload only documents you may use for this purpose and, where possible, no personal data of third parties.

Revising the script. You review the draft of an AI agent before approval; you change scenes yourself or have the script rewritten with notes. These notes are transmitted to the model together with the script and stored with the video. Please do not enter personal data of third parties there.

5. AI processing: currently Anthropic in the USA, in future primarily Google Cloud in the EU

Exploration, script, translation, summarising knowledge sources and review are performed by language models from Anthropic’s Claude model family. We can use them on two routes: through Google Cloud Vertex AI in the European Union or directly through Anthropic’s Claude API (contracting party Anthropic Ireland, Limited, processing by Anthropic, PBC in the USA). Currently (as of 25 September 2026) all requests to language models go through the Claude API in the USA, because the quota we have requested for Vertex AI has not been approved yet. Once it is approved, we use Vertex AI primarily and the Claude API only as a fallback; we will update this policy then. Both routes are described here.

Language models via Vertex AI. Through Google Cloud Vertex AI we use the same Claude models for the same purposes. The contracting party and processor is Google Cloud EMEA Limited, 70 Sir John Rogerson’s Quay, Dublin 2, Ireland, under the Google Cloud Data Processing Addendum. Processing is pinned to a region in the European Union. Google does not use inputs and outputs to train models and does not store them permanently after processing; a short-lived cache for recurring inputs (prompt cache) is discarded after at most one hour. When used through Vertex AI, Anthropic, PBC has no access to inputs or outputs under Google’s contractual commitments; to that extent Anthropic is not a sub-processor.

Language models via Anthropic’s Claude API (currently, later as a fallback). While the quota we have requested for Vertex AI is still pending approval, and afterwards as a fallback when Vertex AI is unavailable, we use the Claude models (exploring your software from screenshots and page structure, summarising knowledge sources, drafting scripts and plans, translation, text checks) directly through Anthropic’s Claude API. The contracting party and processor (Art. 28 GDPR) is then Anthropic Ireland, Limited, 6th Floor, South Bank House, Barrow Street, Dublin 4, D04 TR29, Ireland, under Anthropic’s Commercial Terms, which incorporate Anthropic’s Data Processing Addendum with the EU standard contractual clauses; processing is carried out by its affiliate Anthropic, PBC, 548 Market Street, PMB 90375, San Francisco, CA 94104, USA, as a further processor. Processing takes place in the USA (third country, section 13). Under the Commercial Terms, Anthropic may not train models on content from the use of its services. Anthropic retains inputs and outputs for up to 30 days for abuse and safety review and deletes them afterwards. Excepted are inputs and outputs that Anthropic’s automated systems flag as violating its Usage Policy: they are retained for up to two years and the classification scores for up to seven years. Anthropic also retains data where required by law. The software decides per job which route is used; both routes receive the same data and the same restrictions, in particular never credentials.

What is transmitted (on both routes): page texts and page structure of the explored software (with the placeholders from section 4) and screenshots (blacked out as described in section 4), your project description, your notes for revising the script, the script, your brand rules and the pronunciation dictionary, for Excel files the structure, formulas and texts (numbers replaced by their kind), and from additional knowledge sources the texts of the fetched websites, the text sections of confirmed documents and your notes. Not transmitted are credentials, complete e-mails from the project mail address, uploaded files as such and your billing details.

Speech synthesis. The spoken texts of the script are converted to audio with Google Cloud Text-to-Speech (“Chirp 3 HD” voices); here too Google Cloud EMEA Limited is the processor. Requests go through the service’s EU endpoint; according to Google, the data stays within Europe while stored and processed. Only the texts to be spoken and the voice settings are transmitted. We cache generated audio files under a hash of the text to avoid repeated requests.

Recording and rendering. The automated browser and the video generation run as a job on Google Cloud Run in a region in the European Union. Each job runs in its own container that is discarded afterwards; results are uploaded exclusively to our storage (section 7). We keep operational job logs (phase, progress, errors, without credentials and without page content) for 90 days.

Usage data. For each job we record time, model, number of tokens, characters of speech synthesis, compute time and storage use and the resulting costs in order to bill credits and check costs. This usage data contains no content and is part of our accounting records (section 14).

No automated decision-making. Klickdemo makes no decisions with legal effect on individuals (Art. 22 GDPR). Every script is approved by you before a video is produced.

Transparency under the AI Act. Videos from Klickdemo contain synthetic speech. We mark them in a machine-readable way in the metadata of the video file and in a companion file, show the “AI voice” notice in the player and offer a default line in the outro (Art. 50(2) AI Act, applicable since 2 August 2026). The legal basis for AI processing is performance of the contract (Art. 6(1)(b) GDPR) or your instruction as controller (Art. 28 GDPR).

6. Processors at a glance

The service providers named here process data only on our instructions and are bound by contracts under Art. 28 GDPR. The full list with purpose, location and transfer basis is available at klickdemo.com/en/subprocessors and in Annex 1 of our data processing agreement; we announce changes there at least 30 days in advance.

7. Storage, delivery and mail receipt: Cloudflare

Videos, screenshots, audio files and uploaded files are stored in the object storage Cloudflare R2 of Cloudflare, Inc., 101 Townsend Street, San Francisco, CA 94107, USA (Cloudflare Germany GmbH, Rosental 7, 80331 Munich, as European contact), under the Cloudflare Data Processing Addendum. The storage is set to the “EU” jurisdiction: data is stored exclusively in data centres in the European Union. Cloudflare encrypts stored objects at rest by default. The buckets are not public; the application and the worker access them through short-lived signed URLs. Cloudflare delivers released videos to viewers through its network; the Cloudflare location closest to the viewer processes the viewer’s IP address for delivery and attack mitigation. Cloudflare also manages the DNS records of our domains klickdemo.com and klickdemo-mail.com and receives the messages to the project mail addresses at klickdemo-mail.com (section 4) with Cloudflare Email Routing: a Cloudflare Worker passes each message immediately over an encrypted connection to our server. Cloudflare does not store the content of the messages; the Email Routing activity log contains sender, recipient, time and delivery status. Videos that have not been viewed or downloaded for twelve months are moved within Cloudflare R2 to the cheaper storage class “Infrequent Access”; location, encryption and access stay the same. For possible access from the USA, the adequacy decision for the EU-US Data Privacy Framework (Cloudflare, Inc. is certified) and the EU standard contractual clauses apply.

8. Embed player and view counts

The player you use to embed videos on your website and the share page of a video set no cookies and store nothing in the viewer’s browser (no localStorage either); the unlock token of a password-protected share link only lives in the memory of the open page. The player loads videos, captions and poster images through short-lived signed addresses from our object storage (Section 7); it embeds no other services. During playback we receive the viewer’s IP address, the browser identifier, the origin of the embedding page (such as https://www.example.com) and these events: start of playback, reaching 25, 50 and 75 percent, end and language switch, each with the language of the video. So that a viewer is not counted twice within a day, we compute a hash (HMAC-SHA256) of IP address and browser identifier with a random daily key; we do not store the IP address or the browser identifier themselves. The daily key and the hashes are deleted when the day (UTC) is over, after which no value can be linked to a person. To mitigate abuse, the server counts requests per IP address for one minute in memory. Only totals per video, day and language are stored (starts, distinct viewers on that day, quartiles reached, ends, language switches), plus starts per embedding domain. Individual viewers cannot be identified from this. We need the domain to enforce the embed domains and view allowances of each plan and to show you where your video is played.

For your website you are the controller; we process viewer data on your behalf. Please name Klickdemo in your privacy policy as a processor for the delivery of embedded videos; consent from your visitors is not required for the player because no information is stored on or read from their devices (Section 25 TDDDG). We use the same player on our own website (Art. 6(1)(f) GDPR).

9. Payment processing, invoices, fraud prevention

Payments are processed by Stripe Payments Europe, Ltd., 1 Grand Canal Street Lower, Dublin 2, Ireland (“Stripe”), as an independent controller. Stripe receives name, company, e-mail address, billing address, VAT identification number where applicable, amount and payment method; we do not receive full card details, only confirmation, amount, time, the last digits and a technical identifier of the payment method (Art. 6(1)(b) GDPR). Stripe issues invoices and credit notes on our behalf; Stripe Tax calculates VAT from your billing address and VAT identification number. Stripe screens payments for fraud (Stripe Radar) and may process device data and the IP address for this; Stripe’s privacy policy: https://stripe.com/privacy. Data may be transferred to Stripe, Inc. in the USA (Data Privacy Framework, standard contractual clauses).

To prevent multiple accounts and chargeback fraud we store the technical identifier of the payment method and match it against other accounts at the first purchase (Art. 6(1)(f) GDPR). A block after a chargeback is reviewed by a person before it remains in place for more than 30 days. If we store a card for automatic top-up, this happens only with your consent in the purchase flow; you can remove it at any time in the customer portal. For purchase on invoice (annual contract) we additionally process payment receipt and dunning status.

10. Partner programme and Castfold

Partner programme. If your account takes part in the partner programme, we store your code, the referred accounts (as identifiers, without their names for you), the commission bookings and your payout details (name or company, country, tax status, VAT identification number, for partners in the USA and Canada optionally the tax number of the business – EIN or Business Number, never a social security number –, language) as well as the version and time of your consent to the self-billing procedure including the declaration of tax residence. For payouts Stripe sets up a payout account (Stripe Connect) and verifies identity and bank details as an independent controller under anti-money-laundering rules; date of birth, address, phone number, ID or social security details and bank details are entered only at Stripe, and we do not receive them. We transmit to Stripe your name or company name, e-mail address, country, language, the payout currency of your country, a pre-worded business description (“referral partner of Klickdemo”), for partners in the USA and Canada the industry code 7311 (advertising), and the payout amounts (Art. 6(1)(b) GDPR). From Stripe we receive the status of the payout account (open requirements, enabled capabilities) and, if a payout to your bank fails, its error code. If you end your participation, we have the payout account closed at Stripe once nothing is left to pay out. We retain credit notes for eight years.

Publishing through Castfold. Castfold is another product of Droidtech e.K., that is, of the same controller. If you want to publish a video on YouTube, you store an access key generated in Castfold in Klickdemo; we store it encrypted and you can revoke it at any time. On your instruction we transfer the video file, title, description and chapters to Castfold, which uploads them to the YouTube channel you selected. From the transfer to YouTube onwards, Google LLC (or Google Ireland Limited) processes the data under its own responsibility according to the YouTube Terms of Service and the Google Privacy Policy. Castfold’s own privacy policy at castfold.com applies to Castfold.

11. Sending and receiving e-mail

We send sign-in codes, invitations, confirmations and notifications through the e-mail service of Neue Medien Münnich, owner René Münnich (ALL-INKL.COM), Hauptstraße 68, 02742 Friedersdorf, Germany (processor, servers in Germany, TLS connection). Our mailbox mail@klickdemo.com is also located there. We receive messages to the project mail addresses through Cloudflare (section 7). E-mail address, name, subject and message content are processed. Our e-mails contain no tracking pixels.

12. Support and contracts

We handle requests to mail@klickdemo.com, complaints and cancellations on the basis of the contract (Art. 6(1)(b) GDPR). When you accept the data processing agreement in the application or place an order, we store the version, time, acting person, IP address and a checksum of the document as evidence (Art. 6(1)(c) and (f) GDPR).

13. Recipients and international transfers

Recipients of personal data are the processors named above (Host Europe, ALL-INKL.COM, Google Cloud EMEA Limited, Anthropic Ireland, Limited with processing by Anthropic, PBC, Cloudflare, Inc., Stripe as invoice issuer), Stripe as an independent controller for payments, Google Ireland Limited and Google LLC for the consent-based website services, and Google LLC/YouTube when you instruct a publication. Authorities receive data only where we are legally obliged.

Application, database and object storage are located in Germany and the European Union. Currently the requests to language models (Anthropic, USA, below) go to a third country. For the services of the Google, Cloudflare and Stripe groups, access from the USA (for example by support or security functions) cannot be fully excluded. The basis for such a transfer is the European Commission’s adequacy decision for the EU-US Data Privacy Framework (Art. 45 GDPR), under which Google LLC, Cloudflare, Inc. and Stripe, Inc. are certified, and additionally the EU standard contractual clauses (Art. 46(2)(c) GDPR).

Anthropic (USA). A transfer to a third country takes place insofar as AI processing runs through Anthropic’s Claude API (contracting party Anthropic Ireland, Limited, processing by Anthropic, PBC; section 5; currently for all requests to language models): the data named there (page texts and page structures with placeholders, screenshots, project description, notes for revising the script, script, brand rules, pronunciation dictionary, structure, formulas and texts of Excel files, texts from knowledge sources and notes) is processed in the USA. The basis is the EU standard contractual clauses (Art. 46(2)(c) GDPR), which form part of Anthropic’s Data Processing Addendum; the Data Processing Addendum is incorporated into Anthropic’s Commercial Terms. This transfer does not rely on the adequacy decision for the EU-US Data Privacy Framework. Credentials and billing details are not transmitted.

We provide copies of the clauses on request.

14. Retention

Data	Period
Server logs (website, application, player)	14 days
Salted IP hash for the teaser limit	24 hours
Teaser file, recording and video, company name and e-mail address of the teaser	2 days after being made available (48 hours; without a video after the upload), unless an account is created with the same address
Domain of a teaser’s e-mail address (part after the @)	180 days after the request
Sign-in codes	valid until used, at most a few minutes; the entry for limiting failed attempts (hash of the code, e-mail address, IP address) 48 hours
Sessions	until sign-out or expiry, then at most one day
Change of the sign-in address (previous and new address)	completed ones 90 days, unfinished or rejected ones 48 hours
Invitations (e-mail address and role of the invitee)	90 days after acceptance, revocation or expiry
Activity log of the account (who did what and when)	12 months
Delivery log of job notifications (recipient, occasion, without the content of the mail)	90 days after sending
Requests through the contact and the enterprise form	6 months after they were settled, spam 30 days; if the request leads to a quote, like the quote (below)
Demo-account credentials	until the source, project or account is deleted
Messages to the project mail address	30 days after receipt, at the latest with the project
Intermediate files, audit screenshots	no later than 30 days after the run
Raw take of a video version (silent picture track, narration track, action log)	together with that version
Older versions of a video (beyond the current and the previous one per language)	30 days after two newer versions replaced them
Knowledge base, screenshots, scripts, revision notes, files	until the project or account is deleted
Additional knowledge sources (documents, fetched website texts, notes) and their entries in the knowledge base	until the source, the project or the account is deleted
Projects, videos, recordings, knowledge bases and access data after the end of a plan	fully usable for 90 days (embed, share link, download), then embed and share link off; deletion after a further 90 days (notice at the end and 30 and 7 days before; a new plan stops the period)
Projects with their knowledge base (including screenshots) and videos of free accounts without a plan (teaser)	90 days after the last sign-in (notice 30 and 7 days before; a sign-in resets the period)
Videos above the storage allowance when storage credits are not settled	no earlier than 90 days after the first notice, only videos neither viewed nor downloaded for twelve months (notice 30 and 7 days before)
Storage measurement (type and size per file, daily totals per account, last view per video)	until the file, the video or the account is deleted; daily totals per account at most 36 months
Operational job logs (without content)	90 days
Player view counts (totals per video and day)	until the video is deleted
Account data	until the account is deleted
Origin of the registration (campaign details, click identifier, partner code, landing page, referring website)	in the browser only with consent, at most 90 days; with the account until it is deleted; for a teaser without an account like the teaser’s company name and e-mail address
Payment-method identifier against multiple accounts	12 months after account deletion
Event notifications of the payment provider (webhooks)	90 days after processing
Automatic top-up attempts	13 months (invoice and booking as below)
Audit log of the administrators (changes to prices, credit grants, accounts)	24 months; the deletion record of an account 6 years (below)
Invoices, credit notes, billing details, Stripe identifiers	8 years (§ 14b UStG, § 147 AO, § 257 HGB)
Credit ledger, usage data without content	10 years (§ 147 AO, § 257 HGB)
Order confirmations, quotes, consents, DPA acceptances, cancellations, deletion record	6 years (§ 257 HGB, § 147 AO); afterwards address and e-mail address are removed

Statutory retention periods run from the end of the calendar year. At the payment provider we immediately detach all stored payment methods after an account deletion and have the customer object deleted 180 days after the last payment or refund; invoices remain there for the period stated.

Deleting the account. You start the deletion in the application and confirm it with a code sent by e-mail. We then remove projects, credentials, knowledge bases, videos, memberships and access keys without undue delay and anonymise the account. Restoration is not possible; download your videos and export your data beforehand.

15. Your rights

You have the right of access (Art. 15 GDPR), rectification (Art. 16), erasure (Art. 17), restriction of processing (Art. 18), data portability (Art. 20) and to object to processing based on legitimate interests (Art. 21). You may withdraw consent at any time with effect for the future. Send requests to mail@klickdemo.com. You can lodge a complaint with a supervisory authority; the competent authority is the State Commissioner for Data Protection and Freedom of Information of Rhineland-Palatinate, Hintere Bleiche 34, 55116 Mainz, Germany.

You obtain your data export (Art. 20 GDPR) in the application under Settings › Privacy: a ZIP with account, project, script, knowledge-base and usage data and, for the account owner, invoices and credit history; you download the ZIP file directly, we do not store it. Credentials and keys are not included; videos are downloaded individually.

Data subjects from our customers’ projects (for example employees whose data appears in a demo account) should contact the respective customer as controller; we forward requests that reach us to the customer without delay.

16. Changes

We update this policy when features, service providers or the legal situation change. The current version with its date is always at klickdemo.com/en/privacy/.