Privacy Policy

Effective date: 2026-08-05

App: Brief — understand official letters ("the App")

Applies to: the iOS application distributed via the Apple App Store in Germany.

1. Controller

The controller within the meaning of Art. 4(7) GDPR is:

Vasile Piciriga

Odenwaldstr. 1

63820 Elsenfeld, Germany

E-mail: vasilepiciriga@gmail.com

We are established in Germany, so no representative under Art. 27 GDPR arises. For any question

about this policy or about your rights, write to the address above.

2. Our privacy approach in one paragraph

Brief is designed to be privacy-first. Text recognition (OCR) of your document photos happens entirely on your device. Only when you request an AI analysis is a sanitized text excerpt of the recognized text sent — through our own secure backend proxy — to an AI service for processing. Your documents are never stored on our servers after processing, and document content never appears in our logs. The App shows no advertisements. With your permission — and only then — it measures its own advertising through two advertising networks (Section 8); your letters, their text and your replies are never part of that measurement.

3. What data is processed, where, and why

3.1 Document photos and OCR text — on your device only

When you photograph or import an official letter, the image and the OCR text extracted from it are processed and stored locally on your device (in the App's protected storage, covered by iOS data protection/encryption).
Photos and OCR text are not uploaded to us or anyone else as part of scanning. OCR runs fully on-device using Apple's on-device frameworks.
Locally stored documents remain under your control: you can delete individual documents or all App data at any time inside the App (see Section 9).

3.2 Text excerpts sent for AI analysis — via our backend proxy

When you actively request an analysis, the App sends the recognized text (and, where necessary for quality, limited structural metadata such as document type hints and your selected UI language) to our backend.
Before sending, the App removes lines that look like an attempt to manipulate the AI system, and transmits recognised text only — never the photograph or the PDF. It does not attempt to strip names, addresses or reference numbers: doing so would destroy the very information the explanation depends on. Note: because the purpose of the service is to explain your letter, the transmitted text can inherently contain personal data (e.g. your name, address, file numbers).
Our backend acts as a proxy: it forwards the text to our AI provider (Section 6), returns the result to your device, and does not persist document content. Document text is processed transiently in memory and is not written to server-side logs or analytics.
Transmission is encrypted in transit (TLS).

3.3 Anonymous install identifier

The App uses a randomly generated, anonymous install identifier (not linked to your name, email, Apple ID, or device serial). Purposes: enforcing the free tier and the Pro fair-use limit (50 AI analyses per month), abuse prevention, and technical troubleshooting.
The identifier is stored in the device keychain and is destroyed by "Delete all data"; deleting and reinstalling the App also produces a new one.

3.4 Subscription status via Apple

Purchases are handled entirely by Apple (App Store / StoreKit 2). We do not receive your name, address, or payment details from Apple. We receive/verify only subscription entitlement information (e.g. signed transaction/renewal info, product ID, expiry, original transaction ID) to unlock Pro features and prevent fraud.
Apple's own processing is governed by Apple's privacy policy.

3.5 Technical data strictly necessary for operation

When the App communicates with our backend, standard connection data is processed (IP address, timestamps, request status). IP addresses are used only for connection handling and security. We keep no application log of them ourselves; our hosting provider keeps its own edge logs under its terms, and we are not able to state a retention period for those. Document content never appears in logs.

3.6 What we do NOT do

No advertisements inside the App, no analytics of your documents, no fingerprinting, no sale or sharing of personal data for marketing, no profiling of your letters, and we never use your documents to train AI models. The App does contain two advertising-measurement SDKs (Meta, TikTok). With your permission they report that this installation came from one of our advertisements and whether a few generic milestones followed; Section 8 lists exactly what they can receive and what they can never receive. What the AI provider does is governed by our contract with them; until that contract is concluded and published in the subprocessor list, we do not claim more than this.

4. Legal bases (Art. 6 GDPR)

Processing	Legal basis
Providing the core service you request (OCR on device; AI analysis of text you submit; returning results; free-tier/fair-use accounting; subscription entitlement checks)	Art. 6(1)(b) GDPR — performance of a contract / steps at your request
Security, abuse prevention, ensuring technical stability, defense of legal claims	Art. 6(1)(f) GDPR — legitimate interests (secure and reliable operation of the App; interests balanced against your rights, given the minimal data involved)
Sending recognised text to the AI service for analysis	Art. 6(1)(a) and Art. 9(2)(a) GDPR — your explicit consent, requested before the first analysis that leaves your device and withdrawable at any time in Settings
Advertising measurement with the advertising identifier (attributing an installation and generic in-app milestones to our own advertisements via Meta and TikTok)	Art. 6(1)(a) GDPR — your consent, given in the App Tracking Transparency prompt and withdrawable at any time in iOS Settings → Privacy & Security → Tracking

Aggregated advertising measurement that works without the advertising identifier (Apple SKAdNetwork, Meta Aggregated Event Measurement)	Art. 6(1)(f) GDPR — our legitimate interest in knowing whether our advertising works; no identifier, no content, and you may object at any time by declining tracking

5. Special categories of data (Art. 9 GDPR) — important notice

Official letters you choose to scan may contain special categories of personal data within the meaning of Art. 9(1) GDPR — for example health information (medical or insurance letters), data revealing legal proceedings, or other sensitive content.

Such data is processed only at your own initiative: you decide which document to photograph and whether to request an AI analysis. Before the first analysis that would leave your device, the App asks for your explicit agreement, names the AI service that receives the text, and offers an on-device analysis if you decline. We rely on Art. 9(2)(a) GDPR (explicit consent) for special-category data.
Sensitive content is used exclusively to generate the analysis you requested. We never use it for advertising, profiling, model training or any other purpose, and it is not retained server-side after processing (Section 7). What the AI provider does with it is governed by our contract with them, on the same terms stated in Section 3 — we do not claim more than that here.

6. Recipients and processors

We use a small number of service providers as processors under Art. 28 GDPR, bound by data processing agreements:

Processor	Purpose	Location / transfer safeguard
Anthropic PBC (Claude API)	Generating the AI analysis and the reply draft from the text excerpt you submit	United States — a third country. See Section 6.1.
Cloudflare, Inc. (Workers, KV, Durable Objects)	Hosting our backend proxy, the fair-use counters and the entitlement records	Processing at the edge, including outside the EU/EEA. See Section 6.1.

Apple Inc. / Apple Distribution International Ltd.	App distribution, in-app purchases, subscription management	Apple acts largely as an independent controller for App Store transactions

Advertising-measurement partners — independent controllers, not processors. With your App Tracking Transparency permission, Meta Platforms Ireland Limited (Meta App Events SDK; privacy policy) and TikTok Technology Limited (TikTok App Events SDK; privacy policy) receive your device's advertising identifier together with the generic events listed in Section 8, and process them under their own privacy policies and for their own purposes as independent controllers. Neither receives any content of your letters, and neither is involved in the AI analysis or our backend. Both may process the data outside the EU/EEA under the safeguards described in Section 6.1.

We do not share personal data with any other third parties, except where required by law or to establish, exercise, or defend legal claims.

6.1 International transfers

Where a processor processes data outside the EU/EEA, we ensure an adequate level of protection via: an adequacy decision of the European Commission (Art. 45 GDPR — including, for certified US providers, the EU-U.S. Data Privacy Framework), or the EU Standard Contractual Clauses (Art. 46(2)(c) GDPR) with supplementary measures where needed.

For the AI provider this is the mechanism we rely on, and we are not able to tell you today that the standard contractual clauses are already concluded for your data. Until we can, the App offers on-device analysis: it never leaves your phone, and you can decline the cloud analysis at the consent screen and at any time afterwards in Settings.

7. Retention

Server-side: document text is processed transiently and never stored after processing completes. No copies, no backups of document content. Diagnostic logs record the request method, path, status, duration and the number of pages and characters — never document content, and never any value derived from it — and are retained by our hosting provider under its own default retention; Brief configures no longer period and no content-bearing log.
On your device: scanned documents, OCR text, analyses, and drafted replies are stored locally until you delete them (per document, or via "Delete all data" in Settings). Deleting the App removes all local App data.
Subscription/entitlement records: minimal entitlement data is kept for as long as needed to provide Pro and to meet statutory retention duties (e.g. tax/commercial law, §§ 147 AO, 257 HGB, where applicable to transaction records).
Anonymous install identifier and fair-use counters: kept while the App is installed; counters cover rolling monthly windows.

8. Advertising measurement (App Tracking Transparency)

The App shows no advertisements. It does contain two advertising-measurement SDKs — Meta App Events (Meta Platforms) and TikTok App Events (TikTok) — whose only job is to tell those two networks whether an installation of Brief came from an advertisement we paid for, and whether a few generic milestones followed. That is the definition of tracking under Apple's App Tracking Transparency, and the App treats it as such.

What is asked. During onboarding — or, for installations that predate this feature, once on the home screen — the App explains this and then shows iOS's App Tracking Transparency prompt, once. Declining changes nothing about how the App works. You can change your answer at any time in iOS Settings → Privacy & Security → Tracking, and see your current answer in the App under Privacy & security.

What can be sent. The events are a fixed list built into the App: installation, app launch, onboarding completed, the paywall shown (and which of its three reasons), the first successful analysis of a letter (once per installation, carrying no information about the letter), a free trial started, and a first paid subscription period (the plan, Apple's price and the currency). With your permission the device's advertising identifier (IDFA) travels with them.

What can never be sent — because the App has no code path that could send it: your letter photos and PDFs, the recognised text, summaries and explanations, your replies, names, addresses, reference numbers, amounts, dates or any other content of a letter, the sender or the type of a letter, and the anonymous installation identifier used with our own backend.

Without your permission. No advertising identifier is read or sent. iOS blocks the TikTok SDK from contacting TikTok's servers at all; what TikTok then receives is Apple's own SKAdNetwork postback, which reports in aggregate that an installation happened without identifying you or your device. The Meta SDK likewise relies on SKAdNetwork and on Meta's Aggregated Event Measurement, which work without the identifier.

Recipients and legal basis. Meta and TikTok process this data as independent controllers under their own privacy policies (Section 6). With your permission the legal basis is your consent (Art. 6(1)(a) GDPR); the aggregated measurement that works without the identifier rests on our legitimate interest in knowing whether our advertising works (Art. 6(1)(f) GDPR), and you may object to it at any time by declining or withdrawing tracking permission in iOS Settings.

App Store label. In addition to the three data types of Section 3, the App Store privacy label declares Device ID, Purchase History and Product Interaction as used for tracking, which matches the PrivacyInfo.xcprivacy in the shipped binary. The share extension contains neither SDK and tracks nothing.

9. Your rights — and how the App maps to them

You have the following rights under GDPR:

Access (Art. 15), rectification (Art. 16), erasure (Art. 17), restriction (Art. 18), data portability (Art. 20), objection to processing based on legitimate interests (Art. 21), and the right not to be subject to automated decision-making with legal effects (Art. 22 — the App's analyses are informational only and produce no legal effects; a human — you — always decides what to do).
In-app self-service: the App includes an Export function (export your documents, analyses, and drafts in a portable format — supports Art. 15/20) and a Delete all data function (removes all local data; since we store no document content server-side, this effectively erases your content — supports Art. 17).
For anything else (e.g. questions about backend-side data such as entitlement records), contact us at vasilepiciriga@gmail.com. We answer within one month (Art. 12(3) GDPR).
Where processing is based on consent, you may withdraw it at any time with effect for the future.

10. Children

The App is not directed at children. It is intended for persons who receive official correspondence, i.e. generally 16 years or older. We do not knowingly process children's data.

11. Data security

We use appropriate technical and organizational measures (Art. 32 GDPR): TLS encryption in transit, iOS on-device encryption at rest, minimal-data architecture (on-device OCR, transient server processing), access controls, and no content logging. No method of transmission is 100% secure; we continuously review our measures.

12. Changes to this policy

We may update this policy when the App or legal requirements change. The current version is always available in the App and at https://brief-legal.pages.dev. For material changes we will inform you in the App. The "Effective date" above shows the latest revision.

13. Contact and supervisory authority

Privacy questions and rights requests: vasilepiciriga@gmail.com
You have the right to lodge a complaint with a supervisory authority (Art. 77 GDPR), in particular in the EU member state of your habitual residence, place of work, or place of the alleged infringement. The supervisory authority competent for Vasile Piciriga is: Bayerisches Landesamt für Datenschutzaufsicht (BayLDA), Postfach 1349, 91504 Ansbach, www.lda.bayern.de.