Privacy Policy

Design Extractor (https://www.design-extractor.com) is run by Magentix Studio UG (haftungsbeschränkt), Kaiserring 14-16, 68161 Mannheim, Germany. Full company details are on the Imprint page. Magentix Studio UG is the controller for all processing described here. For anything privacy-related, email info@design-extractor.com.

The short version: we collect what we need to run an account-based extraction service and nothing more. There are no ads, no third-party ad trackers, and nothing is sold to anyone. Product analytics only runs if you accept it, and it runs on EU servers.

You can browse the site and the gallery without an account. Running extractions requires one.

When you sign up with email and password, we store your email address, your display name, and a hash of your password (never the password itself). You must verify your email address; we store the verification state.

When you sign in with Google or GitHub, we receive your name, email address, and avatar from that provider, plus the technical tokens needed to keep the connection working. Google and GitHub are independent controllers of the sign-in step itself; their own privacy policies apply to it.

For every login session we store a session token together with the IP address and browser user agent it was created from. We use this to secure your account, for example to detect abuse and let sessions be revoked.

Legal basis: performance of the contract with you (Art. 6(1)(b) GDPR); session security data also serves our legitimate interest in keeping accounts safe (Art. 6(1)(f) GDPR).

When you run an extraction, we store the URL you submitted, the visual and styling evidence our crawler captures from that public website (screenshots, CSS data, font lists, the page’s visible text such as its title, headings and navigation labels, and images of the individual components we measured, their original and our rendered copy), and the generated result (the DESIGN.md and token exports). On the Pro tier this covers several pages of the same site. Extractions are linked to your account so you can find them again and so credits are accounted correctly.

The captured evidence is stored in a private storage bucket at Cloudflare, located in the EU. It is never public: we read it back only to show you your result, for example the component images on the result page.

Your account keeps a history of the extractions you paid for: the URL, the tier (lite or pro), the credits it cost, the date, and a reference to the exact result you received. We keep that result for you even if the same site is extracted again later by someone else, so what you paid for does not change underneath you. This history is retained for as long as your account exists and is erased when you delete your account (see “How long we keep data”). It is visible only to you and to us, never to other users.

To generate the result, screenshots and styling data of the extracted site (not your personal data) are sent to an AI model provider acting as our processor, Anthropic, through Vercel’s AI Gateway, which relays the request and keeps only request metadata (model, token counts, cost, timing), never the content. To decide which pages of a site to sample and which category a site belongs to, short excerpts of the site’s text (page titles, headings, link labels and page addresses) are sent the same way to a classification model run by TypeSafe, also acting as our processor. Your name and email are never part of these requests.

Some extractions are curated by us into the public gallery. The gallery shows the extracted site’s design spec and brand name, not any data about the user who ran the extraction.

Legal basis: performance of the contract (Art. 6(1)(b) GDPR); gallery curation rests on our legitimate interest in showcasing what the tool produces (Art. 6(1)(f) GDPR).

The Design Extractor extension for Chrome is a client for your existing account. It has no login of its own: it reuses the session cookie you already have from signing in on the site, and sends that cookie only to design-extractor.com. The extension never sees your password.

What leaves your browser: when you click Extract in the popup, the extension sends the address of the tab you are on, together with the tier you chose (Lite or Pro), to our API. The tab’s title stays on your device and only labels the run in the popup. That starts the same extraction described under “Extractions”, with the same credits and the same result, and the run is recorded in your account as started from the extension. While a run is in progress, the extension asks our API for its status. Nothing is sent before you click, and nothing is sent about any other tab.

What the extension does not do: it never reads or uploads the content of the page. Our servers fetch the public page themselves, exactly as they do for a URL typed on the site. It does not record your browsing history, has no access to pages other than the one you extract, contains no analytics, and talks to no service other than design-extractor.com.

What is stored on your device: the extension keeps the last ten extractions you started from it (address, title, tier, status) and your preferred tier in the extension’s local storage, and a snapshot of your account (email address and credit balance) in session storage so the popup can render before the network answers. Session storage is cleared when Chrome closes; the rest stays until you remove the extension or clear its data. Nothing stored by the extension is synced or sent anywhere.

Signing out on the site, or in the popup, ends the extension’s access: it holds no credentials of its own, so without the site’s session cookie it can only offer you the Sign in button.

Legal basis: performance of the contract with you (Art. 6(1)(b) GDPR). The extension’s local storage is strictly necessary for the function you asked for (§ 25(2) TDDDG).

An extraction takes a few minutes, so the site tells you when a run you started in this browser finishes: a message inside the site, and, only if you allow notifications for design-extractor.com, a system notification from your browser. We ask for that permission only when you click “Notify me when it’s ready”, and you can withdraw it at any time in your browser’s site settings.

The notification is created by your browser from the run’s status, which the site already follows while you wait. We use no push service and send nothing to anyone else for it. To know which runs to follow, the site keeps a short list of the runs you started (their public result ID, the site name, and the start time) in your browser’s local storage for up to six hours.

Legal basis: performance of the contract (Art. 6(1)(b) GDPR); the local list is strictly necessary for the function you use (§ 25(2) TDDDG).

Your account has a credit balance and a ledger of credit grants and spends (what was granted or spent, when, and for which extraction). We keep this to run the credit system and to resolve billing questions.

Credit packs are sold through Paddle as merchant of record. Paddle handles the checkout, your payment details, invoicing, and VAT; we never see your full payment data. Paddle’s privacy policy applies to the checkout. We receive confirmation of the purchase (product, amount, and a transaction reference) so we can credit your account.

Legal basis: performance of the contract (Art. 6(1)(b) GDPR) and our legal bookkeeping obligations (Art. 6(1)(c) GDPR).

Account emails

We send transactional emails you cannot opt out of while you have an account: email verification, password reset, and security notices such as a password change confirmation.

Product updates

If you join the update list, we store your email address, the time and page of signup, and your consent state. This is double opt-in: you get a confirmation email and are only on the list after clicking it. Every email contains an unsubscribe link that always works, and unsubscribing is immediate. We use the list only for product updates and launch announcements, never for anything else, and we never share it.

Delivery for both is handled by Resend as our processor. Resend reports delivery events back to us (delivered, bounced, complaint); a hard bounce or spam complaint automatically unsubscribes you.

Legal basis: contract (Art. 6(1)(b) GDPR) for account emails; consent (Art. 6(1)(a) GDPR) for product updates, withdrawable at any time via the unsubscribe link or by emailing us.

The feedback box sends us your message and, only if you choose to add it, your email address so we can reply. Legal basis: our legitimate interest in improving the product (Art. 6(1)(f) GDPR).

On a result you paid for, you can rate its DESIGN.md from 1 to 5 and, if you like, pick reasons and add a comment. We store the rating with your account and the result, never show it to other users, and use it to improve our extractions. Legal basis: our legitimate interest in improving the product (Art. 6(1)(f) GDPR).

Consent-based product analytics (PostHog)

Only if you click Accept on the cookie banner, we load PostHog, hosted in the EU, to understand how the product is used: pages viewed, features used, browser and device type, and the domain (not the full address) of URLs submitted for extraction. PostHog does not receive your email address or other directly identifying data. If you decline, PostHog never loads. You can change your choice at any time via “Cookie settings” in the footer.

Consent-free usage counts

Independently of the banner, we count views, copies, and downloads per extraction in our own database. These events contain no name, email, account ID, or IP address, only the event type, a timestamp, and a random session token that is discarded when you close the tab. We use these counts to show, for example, view counts on gallery items and to see which features are used. Legal basis: legitimate interest (Art. 6(1)(f) GDPR); the associated device storage is strictly necessary for accurate counting (§ 25(2) TDDDG).

Hosting analytics (Vercel)

Our hosting provider Vercel provides aggregate, cookie-free web analytics (page views and web vitals). No cross-site tracking, no profiles, nothing stored on your device.

Here is the complete list of what we store on your device:

None of these are advertising or cross-site tracking cookies. The functional entries are strictly necessary for the service you request (§ 25(2) TDDDG); PostHog storage is set only with your consent (§ 25(1) TDDDG). You can withdraw or change your consent at any time via “Cookie settings” in the footer.

We do not sell, rent, or share your data with advertisers or data brokers. The following providers process data on our behalf:

• Vercel (USA, EU-US Data Privacy Framework): hosting of the web app, hosting analytics, and the AI gateway that relays model requests to Anthropic (retains no prompt or response content; keeps request metadata in its logs)
• Railway (USA, Standard Contractual Clauses): hosting of the extraction backend
• Neon (EU): Postgres databases
• Cloudflare (storage in the EU; Cloudflare, Inc., USA, EU-US Data Privacy Framework): private storage of extraction evidence (screenshots, captured page data, component images)
• Upstash (EU): Redis for job queues, caching, and rate limiting
• Resend: email delivery
• PostHog (EU): product analytics, only with your consent
• Anthropic (USA, EU-US Data Privacy Framework): AI processing of extraction evidence
• TypeSafe(reached only through Vercel’s AI Gateway, which relays the request): classification of sampled pages and of the extracted site’s category from short excerpts of its text
• Paddle (UK): merchant of record for credit packs

Where a provider is outside the EU/EEA, transfers rest on an adequacy decision (EU-US Data Privacy Framework) or on the EU Standard Contractual Clauses.

• Account data: for as long as your account exists. You can delete your account yourself on the account page: after you confirm via email, the account, your extraction history, credits, API keys and connected agents are erased immediately, except records we must keep by law (for example billing records, 10 years under German tax law). You can also email us and we will erase it within 30 days. Besides those records we keep the random account ID, with no other details, so that sign-in tokens issued before the deletion stop working.
• Login sessions: until they expire or are revoked.
• Extractions and their captured evidence: while they serve the product (your history, caching, the gallery). Results you paid for are kept for as long as your account exists so your history stays complete. You can ask us to delete extractions tied to your account.
• Product update list: until you unsubscribe or ask for deletion. Unsubscribed entries are kept only as a suppression record so we never email you again.
• Feedback: until processed, then deleted or anonymized.
• Ratings: the comment is kept for as long as your account exists and deleted with it; the score and reasons are kept without any link to your account. Deleting a result from your library keeps its rating, which belongs to the extraction run, not to the library entry.
• Usage counts: kept as aggregate numbers; the session tokens they reference are meaningless after the session ends.

Under the GDPR you can ask us for access to your data, correction, deletion, restriction of processing, and a portable copy, and you can object to processing based on legitimate interest. Where processing rests on consent, you can withdraw it at any time with effect for the future.

You can delete your account yourself on the account page. To exercise any other right, email info@design-extractor.com from the address on your account. We respond within 30 days.

You can also lodge a complaint with a data protection authority. For us, the competent authority is the Landesbeauftragte für den Datenschutz und die Informationsfreiheit Baden-Württemberg.

Design Extractor is not directed at children under the age of 16, and we do not knowingly collect their data. If you believe a child has provided us with data, please contact us and we will delete it.