1. Data Controller
The data controller for the purposes of the EU General Data Protection Regulation (GDPR) is:
John Michael Mullen
Forststraße 1
12163 Berlin, Germany
Email: privacy@thechronicler.app
2. Data We Collect
We collect the following categories of personal data:
- Account information: Email address, display name, and authentication provider (Google, Discord, or Twitch). If you sign in via OAuth, we receive a limited profile from the provider (typically email and name).
- Campaign content: Audio recordings of game sessions, transcripts generated from those recordings, and all AI-extracted data including entities, relationships, quest threads, session recaps, and highlights.
- Usage data: Pages visited, features used, session processing events, and error logs. Page views and performance metrics are collected via Vercel Analytics (cookieless, no user identifiers); a small set of public pages also send an anonymous pageview beacon to a cross-venture analytics system we separately operate, and that same system receives two anonymous milestone signals — that an account was created, and that a first session entered processing. Those two signals carry an event name and a fixed label only: no account identifier, no email, no campaign or session identifier, and no campaign content. See Section 6 and Section 9.
- Newsletter subscription: If you subscribe to the newsletter — which needs no account — we store your email address and a record of your consent. Section 12 sets out exactly what that record contains and why.
- Payment data: Subscription plan and billing status. Payment processing is handled entirely by Stripe — we do not store credit card numbers, CVVs, or bank account details on our servers.
3. How We Use Your Data
- Providing the Service: Processing audio recordings, generating transcripts, extracting entities and relationships, tracking quests, and producing session recaps and highlights.
- Billing: Managing your subscription, processing payments through Stripe, and sending billing-related communications.
- Service improvement: Analysing aggregated, anonymised usage patterns to improve the Service. We do not use your campaign content to train AI models of our own; what the AI providers we send content to may do with it is set out in Section 6. You may opt out of anonymised analytics at any time by contacting us.
- Communication: Sending transactional emails (account verification, subscription confirmations, payment failures) and, if you opt in, product update notifications.
4. Legal Basis for Processing (GDPR Article 6)
We process your personal data on the following legal bases:
- Contract performance (Art. 6(1)(b)): Processing necessary to provide the Service you have signed up for, including audio transcription, AI extraction, and content generation.
- Legitimate interests (Art. 6(1)(f)): Maintaining service security, preventing fraud, and improving the Service based on aggregated usage data.
- Consent (Art. 6(1)(a)): Where required, such as for optional marketing communications or use of campaign content for model improvement. You may withdraw consent at any time.
5. Data Retention
- Account data: Retained while your account is active. Requesting deletion starts a 7-day hold. We lock the account the moment you ask: you and your players lose access to your campaigns and no session processing runs. Your right of access and portability is not suspended — exporting your data from Account → Settings keeps working for the whole 7 days, and after that there is nothing left to export, so take a copy before the hold runs out. We email you a one-click link to cancel; cancel within the 7 days and nothing is lost. If you do not cancel, we permanently delete your campaigns and everything in them — sessions, transcripts, audio, wiki, quests, timeline — along with your login and profile. A job runs once a day to do this, so the deletion happens on, or within about a day after, the date we gave you. Two things are kept on purpose afterwards: the billing records described below, and one record of the deletion itself (the email address on the account, the date you asked and the date we finished), which we keep for as long as we need to be able to show that the deletion actually happened. See Backups below for why a copy can survive a little longer than that.
- Shared campaigns: A campaign belongs to the GM who created it. If that GM deletes their account, the campaign and everything in it goes with them, including character sheets, inventory entries and feedback their players added — we do not keep a player's copy and we do not email the players a warning, so a GM who deletes should tell their table first. If you are a player and you delete your own account, the GM's campaign is untouched; only your membership and your own login are removed.
- Audio recordings: Kept in storage restricted to your account for 7 days after processing finishes — the same window on every plan — and then automatically deleted. During those 7 days you can download the recording from its session page over a private link that expires within the hour, and about 48 hours before the deletion date we email you a reminder with a link to that page; we never attach audio to an email. Recordings made before 14 November 2026 predate this rule: those files are still stored, and we will write to the GMs who own them, with a download window, before any of them is deleted. Transcripts, recaps and the wiki are not affected and are retained as part of your campaign for the life of your account. You can request earlier deletion of a recording's audio file at any time by contacting us (Section 11).
- Billing data: Invoices and the accounting records behind them are kept for 10 years, because German tax and commercial law requires it (§ 147 AO, § 257 HGB). This applies after you delete your account as well, and it is the one thing we cannot erase on request — the GDPR allows this where the law obliges us to keep the records (Art. 17(3)(b)). Nothing in those records is used for anything other than tax and accounting.
- Usage logs: Retained for up to 90 days for debugging and service reliability purposes.
- Backups: Our database host keeps rolling encrypted backups so the Service can be restored after a failure. Deleted data can still sit inside a backup until that backup is rotated out of the host's retention window, after which it is gone. We never restore a deleted account from a backup. If you want the exact current backup window in writing, ask us at privacy@thechronicler.app.
6. Data Sharing
We share your data with the following third-party processors, solely for the purposes described:
- Supabase (database hosting and authentication) — your account data, campaign content, and transcripts are stored on Supabase infrastructure.
- OpenAI (AI processing and image generation) — transcript segments are sent to OpenAI APIs for entity extraction, relationship detection, recap generation, and other AI features, and AI-generated artwork (such as entity portraits) is produced by OpenAI's image model from prompts built out of your campaign content. Audio is not sent to OpenAI. Text you paste into our free public recap generator (/tools/session-recap) is also sent to OpenAI to produce the recap; we do not store that text or the recap, and neither is linked to an account.
- Deepgram (audio transcription) — uploaded audio recordings are sent to Deepgram for speech-to-text processing.
- Stripe (payment processing) — billing and subscription data is processed by Stripe. Stripe's privacy policy governs their handling of payment data.
- Resend (email delivery) — email addresses are shared with Resend solely for the purpose of sending service-related emails and, where you have confirmed a subscription, the newsletter (Section 12).
- Sentry (Functional Software, Inc.) (error monitoring) — technical fault data is sent to Sentry as described in Section 7 below. Based in the United States; transfers are governed by Standard Contractual Clauses (SCCs).
- Vercel (hosting and page analytics) — the Service is hosted on Vercel's infrastructure, so all request data passes through it. Vercel also provides Vercel Analytics, a cookieless page-view and performance metric used in place of a conventional analytics tracker — no cookies or persistent identifiers are set. Based in the United States; transfers are governed by SCCs.
- Inngest (background job orchestration) — session and campaign processing (transcription, AI extraction, recap and handout generation) runs as a sequence of background jobs managed by Inngest; data produced at each pipeline step passes through, and is briefly retained in, Inngest's execution history. Based in the United States; transfers are governed by SCCs.
- fal.ai (alternative AI image provider) — disclosed ahead of use. AI image generation currently runs through OpenAI (above); no part of the Service sends data to fal.ai today. If we switch or add fal.ai as an image provider, image prompts built from your campaign content would be sent there. Based in the United States; transfers are governed by SCCs.
- Discord (optional notifications and session capture) — if you connect a Discord webhook to a campaign, we post notifications to it when a recap or handout is ready; those notifications contain the campaign and session titles and a link. If you use our Discord commands or recording integration, Discord also processes the command interaction and the voice audio it captures. Discord is additionally one of the sign-in providers named in Section 2. Based in the United States; transfers are governed by SCCs.
- Craig (craig.horse — Discord session recording) — if you import a Craig recording, we fetch the recording's audio and transcript from Craig's servers on your instruction. Craig is an independent, open-source Discord recording service and is the controller of recordings you make with it; we act on the recording link you give us.
- The System OS Collector (the-system-os.vercel.app) — on public, unauthenticated pages we send an anonymous, cookieless pageview/milestone beacon (page path, referrer, and UTM campaign tags only — no cookies, no user identifiers) to this analytics system, which we also operate across our other products, to understand how visitors find the Service. It also records which buttons and links on the page were used and roughly how long the page was open — without any identifier that could tie them to a person, and we suppress the beacon entirely when your browser sends a Do Not Track signal. Separately, our servers send two anonymous milestone signals — account created, and first session entered processing — which carry an event name and a fixed label only, with no account, campaign or session identifier and no campaign content. See Section 9.
We do not sell your personal data to third parties. We do not share your data with advertisers.
AI model training. We do not use your campaign content — audio, transcripts, recaps, wiki text or character sheets — to train any AI model of our own, and we do not sell or license it to anyone else for model training. Neither the transcription provider (Deepgram) nor the AI provider (OpenAI) we send your content to uses it to train their models:
- OpenAI (transcript text, prompts built from your campaign content) — OpenAI's published API terms state that data sent to the API is not used to train or improve OpenAI's models unless the customer opts in. We have not opted in, and we have not joined any data-sharing programme OpenAI offers. OpenAI may keep API content for up to 30 days for abuse monitoring, during which authorised OpenAI staff may review content that its automated checks flag.
- Deepgram (your uploaded audio) — Deepgram's published terms place API requests in its Model Improvement Program by default, under which Deepgram keeps parts of the audio and transcript to improve its own speech models. We set Deepgram's per-request opt-out on every request we send, so audio you upload is not used by Deepgram in that way.
Recognising who spoke. Separating the voices inside one recording is done by Deepgram as part of transcription, so a recap can say who did what. Carrying a speaker's identity from one session into the next is a separate, optional step: for a player it is off by default, only that player can switch it on for themselves (from the player portal), and they can switch it off again at any time. The GM's own speaker row is always carried forward, because the Service cannot produce a usable recap without identifying the GM from one session to the next. We create and store no voiceprint, voice embedding or other biometric template, and this step does not compare voices at all — it works only by reusing a speaker label the GM has already confirmed.
7. Error Monitoring
Error monitoring. We use Sentry (Functional Software, Inc.) as a processor to detect and diagnose technical faults. When an error occurs, technical data — error type, stack trace, browser and device type, page URL, and timestamp — is transmitted to Sentry. This is done on the basis of our legitimate interest in a secure and functional service (Art. 6(1)(f) GDPR). We have configured Sentry not to transmit personal identifiers, and a data processing agreement is in place.
7a. Preventing Abuse of the Free Account
Processing a session costs us real money, and the free account exists so a GM can try the Chronicler without paying for it. To keep that possible, before a free account's session is processed we check that the email address on the account has been confirmed and is not a throwaway inbox. A paid account is never checked this way. These checks never remove anything: a GM asked to confirm their address keeps full access to every session, recap, wiki article and handout already made, and the one step that clears the check is named on screen.
A daily job also looks for signs that one person is opening new free accounts instead of using one — the same recording uploaded under two accounts (matched on its transcript, or on length and file size together), an account that used its whole free allowance and went quiet just before a new one appeared, and repeated recordings over the per-session length limit. It reads only data the Service already holds: session timings, recording length and file size, and the dates accounts were created. The one extra signal is the network a new account first signed in from, and we do not keep that network — we keep a one-way keyed hash of the IP address (HMAC-SHA256, truncated), never the address itself, at most once per account and only while that account is less than a week old, and the same job deletes those hashes after 72 hours. We do not fingerprint your device or browser — no canvas, font or hardware probing of any kind — and this guard writes no raw IP address to any table or log line.
The legal basis is our legitimate interest in preventing abuse of a free service and keeping it affordable to offer (Art. 6(1)(f) GDPR). The strongest thing that can follow is that new processing on a free account is paused pending verification; a person reviews and lifts that, it never applies to a paying account, and it is not a decision based solely on automated processing within the meaning of Art. 22 GDPR. Your campaigns, recaps and wiki are never deleted or hidden by it. If you think the guard has you wrong, write to privacy@thechronicler.app — that is also how you exercise your right to object under Art. 21 (Section 8).
8. Your Rights (GDPR)
Under the General Data Protection Regulation, you have the following rights:
- Right of access (Art. 15): You may request a copy of all personal data we hold about you.
- Right to rectification (Art. 16): You may request correction of inaccurate personal data.
- Right to erasure (Art. 17): You may request deletion of your personal data (“right to be forgotten”). The fastest route is Account → Settings → Delete account, which runs the 7-day hold described in Section 5. You can also write to us at the address below. The hold exists so that a deletion nobody meant to start can be undone; the billing records named in Section 5 are the only data we keep against your wishes, and only because the law requires it (Art. 17(3)(b)).
- Right to data portability (Art. 20): You may request your data in a structured, machine-readable format.
- Right to object (Art. 21): You may object to processing based on legitimate interests.
- Right to lodge a complaint: You have the right to lodge a complaint with a supervisory authority. For users in Berlin, Germany, the relevant authority is the Berliner Beauftragte für Datenschutz und Informationsfreiheit.
To exercise any of these rights, contact us at privacy@thechronicler.app. We will respond within 30 days as required by the GDPR.
9. Cookies
The Chronicler uses a single session cookie for authentication purposes. This cookie is strictly necessary for the Service to function and does not require consent under GDPR Article 5(3) / ePrivacy Directive Article 5(3). We do not use tracking cookies, advertising cookies, or third-party advertising trackers. On our public pages, we record cookieless page counts (page path, referrer, and UTM campaign tags only — no cookies, no user identifiers) via Vercel Analytics and, for a subset of public pages, via a cross-origin beacon sent to the-system-os.vercel.app — a cross-venture analytics system that we also operate, not a first-party request to thechronicler.app. That browser beacon is suppressed when your browser sends a Do Not Track signal. It also records which buttons and links on the page were used and roughly how long the page was open — without any identifier that could tie them to a person. Our servers also send the same system two anonymous milestone signals from inside the signed-in Service (account created, first session entered processing); these are server-to-server, set no cookie, and carry an event name and a fixed label only — no account, campaign or session identifier and no campaign content. No cookies or user identifiers are included in any of these cases. See Section 6.
10. International Transfers
Your data is processed by service providers based in the United States (OpenAI, Deepgram, Stripe, Supabase, Resend, Sentry, Vercel, Vercel Analytics, Upstash, Inngest, Discord, Craig, and — if enabled — fal.ai). This also includes The System OS Collector (Section 6), which runs on the same US-based hosting infrastructure. These transfers are conducted under Standard Contractual Clauses (SCCs) approved by the European Commission, ensuring an adequate level of data protection as required by GDPR Chapter V. Where available, we utilise additional safeguards such as encryption in transit and at rest.
11. Contact / Data Subject Requests
For any questions about this Privacy Policy, or to submit a data subject access request, please contact:
Email: privacy@thechronicler.app
Postal: Forststraße 1, 12163 Berlin, Germany
We will acknowledge your request within 72 hours and provide a substantive response within 30 days.
12. Newsletter (Double Opt-In)
You can subscribe to our newsletter from the home page without holding an account. We use the double opt-in procedure required in Germany: after you submit your address we send one confirmation email, and nothing further is sent unless you click the confirmation link. If you never confirm, the address is never used for the newsletter.
- What we store: your email address, where on the site you subscribed, the exact consent wording you were shown and its version, the time of your request and of your confirmation, and the IP addresses those two steps came from.
- Why the IP addresses: German case law requires the sender of a newsletter to be able to prove that the holder of an address consented. These two records are kept solely as that evidence (Art. 6(1)(f) GDPR, our legitimate interest in defending against unsolicited-email claims).
- Legal basis for sending: your consent, Art. 6(1)(a) GDPR and § 7(2) no. 2 UWG.
- Content: product news and occasional tips for running a game. We do not track opens or clicks, we do not profile subscribers, and we never pass the list to anyone.
- Processor: the emails are delivered by Resend (Section 6), under the same safeguards as our other transactional mail.
- Withdrawal: every newsletter email carries an unsubscribe link that works in one click, with no login. You can also write to the address in Section 11. Withdrawing does not affect anything we sent before.
- Retention: your address and the consent record are kept for as long as you are subscribed. After you unsubscribe, the record of consent and withdrawal is kept for the duration of the statutory limitation period so that we can evidence that the subscription was lawful, and is not used for any other purpose.
13. Updates to this Policy
We may update this Privacy Policy from time to time. Material changes will be communicated via email or a prominent notice within the Service at least 30 days before they take effect. The “Last updated” date at the top of this page reflects the most recent revision.