Privacy Policy

Last Updated: October 14, 2026

Daniel Sticker ("we", "us", "our") operates the Cadenza service (the "Service") as a paid service. This Privacy Policy explains how we collect, use, and protect your personal data in accordance with the General Data Protection Regulation (GDPR) and applicable German data protection law. The data-subject rights described below apply in full to every natural person whose personal data we process.

I. Information We Collect.

When you create an account, we collect your name, email address, and authentication provider (such as Google or GitHub). When you subscribe, we collect the billing details needed to process your payment, handled by our payment processors as described below. If you allow browser analytics, we collect limited usage metadata such as normalized page categories, feature names, and coarse outcome categories. We do not send page query strings, form contents, message or draft text, names, email addresses, profile URLs, search terms, or raw errors to PostHog. PostHog derives a coarse location from the IP address and then discards it. During signup, if you allow browser analytics, your browser sends first-party attribution data to our Convex backend, where we store your first-touch referral or campaign category. When you connect your own LinkedIn account, we process data from that account on your instruction to provide the features you request. This can include storing normalized messages, posts, and conversation context for the Writing Profile and, when you request cleanup features, thread evidence described in Section VII below. When you use the support chat, we collect the messages and any attachments you send.

II. How We Use Your Information.

We use your information to provide and maintain your account, process your subscription and payments, send transactional emails (such as verification, password reset, and support replies), send service-related communications that are necessary to operate your account (for example security notices and material changes to the Service), and improve the Service through aggregated, pseudonymized analytics. We use connected-account data for the feature you request: for example, to draft or deliver an action, maintain your Writing Profile, or preserve the reviewed thread evidence needed to reproduce and improve a cleanup result. We store and analyse your tool calls and their results, including the LinkedIn content they retrieve, to provide support and improve the Service. We do not sell, share, or rent your personal data to third parties for marketing purposes, and we do not sell or resell data from your connected account. Any optional, promotional communications are sent only where you have given consent, and you can withdraw that consent at any time.

III. Legal Basis for Processing.

We process your data on the following legal bases under the GDPR: contract performance (Art. 6(1)(b)) to provide your account, your subscription, and the Service, legitimate interest (Art. 6(1)(f)) to improve the Service and ensure security where that basis is applicable, compliance with a legal obligation (Art. 6(1)(c)) where we are required to retain data, for example the billing and invoice records we must keep for tax and accounting purposes, and consent (Art. 6(1)(a)) where explicitly given, including optional browser analytics and optional communications.

IV. Third-Party Processors.

We use the following GDPR-compliant third-party services to operate the Service: Convex (EU, Ireland) for database and backend infrastructure, Cloudflare (Workers) for hosting and deployment, Unipile for LinkedIn account access, Resend for transactional email delivery, Autumn and Stripe for billing and subscription management, OpenRouter for AI features including server-side draft generation and Writing Profile training, and PostHog for product analytics. Each of these processors maintains appropriate technical and organizational measures to protect your personal data. Section VI describes how we handle payment data, and Section VII describes how connected-account data is handled.

PostHog browser analytics is disabled until you choose to allow it. Session replay, automatic click capture, advertising attribution, surveys, and automatic exception capture are disabled for browser analytics. Your browser choice applies to browser analytics on that browser and device.

For authenticated MCP calls, we process tool names, timings, coded outcomes and pseudonymous account or diagnostic identifiers to detect faults and maintain service reliability and security, based on our legitimate interest (Art. 6(1)(f)). Technical analytics and error-monitoring providers receive this metadata and restricted technical stack locations, never tool inputs, results or raw error messages. This server-side processing does not depend on browser analytics consent.

When a request to one of our webhook endpoints fails authentication, we record its IP address, user agent and request header names to investigate misuse, based on our legitimate interest in the security of the Service (Art. 6(1)(f)), and delete them after about 30 days.

V. International Data Transfers.

Some of our processors operate outside the European Union, including in the United States (for example Cloudflare, Stripe, Resend, and PostHog). Where personal data is transferred to a processor outside the EU or the European Economic Area, the transfer is covered by appropriate safeguards under the GDPR, in particular the European Commission's Standard Contractual Clauses together with any additional measures required to protect your data. You can request more detail about these safeguards using the contact details below.

VI. Payment Processing.

Subscription billing for the Service runs through Autumn and Stripe. Stripe processes your card and payment details directly as a payment processor. We do not store your full card number, and we receive only the limited billing information needed to manage your subscription, such as the status of your payment and a reference to the transaction. Stripe processes payment data in accordance with its own privacy terms and applicable payment-industry standards.

VII. Connected-Account (LinkedIn) Data.

The Service lets you connect and operate your own LinkedIn account to draft and send outreach. We access your connected LinkedIn account exclusively through our integration partner, Unipile (Unipile SAS), acting as our subprocessor for that access. We connect to and access your account only on your instruction and on your behalf, to provide the features you request. We access only the data and actions your own authenticated account is already entitled to, and we do not circumvent any access control, rate limit, or paywall of any connected platform.

When you connect a LinkedIn account, we choose a country for the connection location from your current request, from your own choice, or from your browser's time zone when it points to a single country. We send that country to our integration partner as the requested connection location and store it with your connected account for connection maintenance and recovery. The integration partner may use another location when the requested one is unavailable. We do not send your IP address to the integration partner for this purpose.

To the extent we process data from your connected account on your behalf, you act as the controller and we act as your processor within the meaning of the GDPR, and we process that data only in accordance with your instructions and applicable law. We do not build, hold, or maintain an independent database of third-party LinkedIn profiles, and we do not resell data from your connected account. Disconnecting stops new provider access, but records already retained for a feature remain subject to that feature's retention and deletion rules below.

To provide the features you use, we store the data from your connected account that those features need, such as messages, posts, drafts, and the conversation context around them, including content written by the people you correspond with. We use it to provide and improve those features, for example to train your Writing Profile. Writing Profile learning is enabled by default, and you can pause it in settings at any time.

When you use or request a conversation-cleanup feature, we may retain a cleanup snapshot of each provider-visible thread that the feature fully reviews. The snapshot is a normalized, versioned representation that may include provider item identifiers, message directions, timestamps, message content, and messages written by other participants. We use this evidence to reproduce cleanup results, investigate problems, and develop future cleanup insights. We do not treat cleanup evidence as a complete copy of your inbox or as the source of truth for threads the feature did not fully review. A general inbox mirror or full-text search feature would require a separate product and privacy review. On your explicit approval or restore request, the Service submits a request through Unipile to archive or unarchive the selected LinkedIn conversations and records the outcome.

Cleanup snapshots are kept until you delete your account; they do not expire on a time-based schedule. Deleting your account removes them together with your other product data. Content already returned to an external MCP caller is outside Cadenza's ability to recall.

We store cleanup snapshots in Cadenza's own EU system of record. We do not duplicate the normalized snapshots or their stored message content into PostHog, Telegram, or our error-event store as product evidence. Cleanup monitoring in those services uses coded outcomes and non-content metadata. Raw provider responses are not cleanup evidence; this cleanup rule does not change how the Service handles provider errors for its existing diagnostic and security purposes. Processors listed in Section IV may process content where needed to provide the requested feature.

VIII. Cookies.

We use essential cookies required for authentication and session management. We also store one first-party cookie recording whether you allowed or declined browser analytics. The cookie also includes a change counter so that withdrawals are respected across tabs. PostHog is not loaded before you allow analytics and its SDK does not store a separate durable consent marker. We do not use advertising cookies. You can reopen Analytics preferences and withdraw your choice at any time; no new browser capture is initiated after withdrawal, while a request already accepted before withdrawal may still finish.

IX. Data Retention.

We keep your personal data while your account is active and only as long as the feature that stores it needs it. Cleanup snapshots remain until you delete your account and have no time-based expiry. Account deletion removes the data features stored from your connected account, subject to records we must retain under applicable law. Some data may be retained in aggregated, pseudonymized form for analytics, or as required by applicable law, including the billing and invoice records we must keep for tax and accounting purposes in connection with your payments.

X. Your Rights.

Under the GDPR, you have the right to access your personal data (Art. 15), rectify inaccurate data (Art. 16), request deletion of your data (Art. 17), restrict processing (Art. 18), data portability (Art. 20), and object to processing (Art. 21). These rights apply to every natural person whose personal data we process. Where we rely on your consent, you also have the right to withdraw it at any time (Art. 7(3)), and withdrawal does not affect the lawfulness of processing carried out before withdrawal. To exercise any of these rights, please contact us using the details below. You also have the right to lodge a complaint with a supervisory authority.

XI. Changes to This Policy.

We may update this Privacy Policy from time to time. If we do this, we will post the changes on this page and will indicate the date these terms were last revised. Your continued use of the Service after the date any such changes become effective constitutes your acceptance of the updated Privacy Policy.

Questions?