Chat Digest logoChat Digest Group chats — in one digest
MagyarEnglish
How it works Pricing Help Log in
⚠ Draft — the parts in [square brackets] and the legal review are still pending.

This is an informal English translation. The Hungarian version is the legally binding one. Magyar

Privacy Notice

Effective from: October 28, 2026 · Last updated: 2026-10-28 (draft)

1. Who handles your data

Data controller (for own-purpose data) and data processor (for the content of conversations): Varga Fábián György, sole proprietor, registered office: 1148 Budapest, Nagy Lajos király útja 32-34. III/1., tax number: 46128737-1-41, sole proprietor registration number: 58369546, e-mail (also for data protection inquiries): varga.fabian.gyorgy@gmail.com, phone: +36 30 405 5780. This notice applies to the Chat Digest service (website and browser extension). We do not currently distribute the desktop application, so it is not covered.

In providing the Service we act in two roles, and in this notice we indicate for each category of data which one:

  • Here we are data processors: the Content — the conversation text you upload or paste, the messages processed from it (including the names and messages of others), the finished summary and the "Keep" marker. We process it on your behalf and on your instructions, solely to make, store, display and delete the summary; we do not use it for our own purposes (we do not train a model on it, do not analyze it, and do not pass it on beyond providing the Service). The detailed rules are in the Data Processing Terms (an annex to the Terms).
  • Here we are data controllers: own-purpose data — the account, the subscription, payment and invoicing, the record of withdrawal/termination, security data (logs, request limiting, bot protection), feedback and customer-service correspondence, the AI cost log (technical data), aggregate statistics (numbers only), data tied to legal obligations, and the cookie and language setting.

2. What data we process

  • [Content - here we are data processors] Only what you hand over: a conversation is processed only when you upload its export, paste its text, or paste a copied text in the browser extension and press the "Get summary" button. Chat Digest processes text that the user hands over manually (copied and pasted by the user), or a conversation export that the user made themselves. The Chrome extension sends the text that you copied and pasted into the side panel (and the group name you enter), only when you press the button. The extension has no access to the Facebook or Messenger pages: it has no permission for them, and no code of it runs on those pages; it does not scroll, does not read any other message, and does not send the page address (URL), the page title or other browser data either. The service does not log in to your social account. Please do not send conversations that contain health data (e.g. allergies, illnesses); the contractual arrangements with the AI provider on this are in progress. Before submitting, by ticking a mandatory checkbox you declare that the content contains no health data or other special-category personal data, and no criminal data; do not submit a conversation that contains such data. [to be finalized by the lawyer] The service may be used only with lawfully accessible content; the mere possibility of exporting or copying does not by itself mean unlimited authorization to further use other people's personal data, confidential communications or copyright works.
  • [Own-purpose data - here we are data controllers] Sign in with Google: if you choose this sign-in method, Google passes your name, your e-mail address and your Google ID to us. Of these we store the e-mail address and the Google ID (not the name), use them to create the account and for sign-in, and keep them until your account is deleted. We do not receive your Google password. Google is an independent data controller in its own sign-in process.
  • [Own-purpose data - here we are data controllers] Account data: e-mail address, password (in non-reversible hash form; with Google sign-in there is no password). The source of the registration (for example which community link you came from) is not part of your account data: we record it only as an anonymous count per source, without a date and without a link to your account. Values that were earlier tied to accounts were aggregated and deleted from the accounts. For this we do not use a cookie, and we do not log your visits.
  • [Own-purpose data - here we are data controllers] Session and device data: the session identifier that keeps you signed in; the browser extension's unique access token (tied to your account), its name, and the time of linking and last use. You can connect the extension to your account from the website with one click: the website issues a device token, which the extension stores in your browser. Your password does not pass through the extension. In your account settings you can see the linked devices and revoke any of them.
  • [Content - here we are data processors] Former connected platforms and groups: the platform (Messenger, Facebook groups) connected with an earlier (0.4.x) version of the extension, the name and technical identifier of the followed conversations/groups, and the marker of the last processed message. The extension's version 0.5.0 creates no new connections. We do not delete these (the group name and the last-message marker) on the 90-day retention period of the summaries; we keep them until your account is deleted, or until you disconnect the platform.
  • [Content - here we are data processors] Summaries: the AI summaries made of the conversations — the gist, to-dos, dates, decisions, open questions, key information, links. These may contain the personal data of other participants and of children (names, dates, institution), if these were mentioned in the conversation. We did not design the service for processing health data, so under the Terms you must not upload any. Because the processing does not filter such data out, it may happen that, despite that commitment, information pointing to health (e.g. an allergy) still gets to the AI provider and into the summary. In that case we treat it like the rest of the content: we do not filter it, we do not use it for any separate purpose, and you can delete the summary at any time.
  • [Own-purpose data - here we are data controllers] Usage log: technical data of the AI processings (time, platform, number of units used; we do not store the name of the group) to track costs. Retention: 12 months, then we delete it. The reason for 12 months is the annual cost accounting and cost tracking (a year-on-year comparison needs a full year of data); the log contains no group name. From this data, per-account usage metrics (for example the number of processings and of units used) are also compiled on the operator (admin) interface; this is metadata and does not contain the content of conversations.
  • [Own-purpose data - here we are data controllers] Feedback: if you send feedback on the site, its text, the page you sent it from, and, if logged in, your e-mail address. Legal basis: our legitimate interest in improving the service and fixing bugs. Retention: 1 year, then we delete it; when your account is deleted, we also delete the text of your feedback. The free text can contain anything you typed in - please do not write conversations, personal data or data about other people in it.
  • [Own-purpose data - here we are data controllers] Abuse protection: to limit the number of requests we store a fingerprint of your IP address made with a secret key for a short time (at most 3 days).
  • [Own-purpose data - here we are data controllers] Bot protection: at registration, Cloudflare Turnstile checks whether a human is using the site. For this, Cloudflare receives your IP address and your browser's characteristics (e.g. User-Agent, TLS fingerprint).
  • [Own-purpose data - here we are data controllers] Invoicing: to issue the invoice, the data shown on the invoice (name, e-mail address, address if needed) go into the Számlázz.hu system. We keep accounting documents for the period required by law.
  • [Own-purpose data - here we are data controllers] Payment: the identifier, amount, time and status of payments. We do not see or store bank card data — it is handled by Barion Payment Zrt.
  • [Content - here we are data processors] What we do NOT store on our server: the raw text of your conversations and the uploaded file. They are in the server's memory only while the summary is being made, and they are passed on to the AI provider named in section 3.

3. Who has access to your data

  • [Content - sub-processor] Anthropic (for customers in the EEA the contracting party is Anthropic Ireland, Limited; the processing may also take place in the United States, subject to Anthropic's data processing terms — the DPA and the transfer safeguards under the standard contractual clauses) — as a sub-processor (taking part in our role as data processor), it processes the text of the messages to make the summary (Claude API). According to Anthropic's public information, data sent through the commercial API is by default not used for model training [to be checked: exact source], and is stored for a limited time under Anthropic's own terms — with exceptions (for example abuse investigation, legal obligations). According to Anthropic's public page the default deletion period is 30 days, and content flagged as violating the rules may be kept for up to 2 years [to be checked]. "Does not train on it" and "does not store it" are two separate statements: the first does not imply the second. [to be checked: exact source and status] Anthropic's data processing terms (Data Processing Addendum, effective 2025-02-24) are part of its Commercial Terms (effective 2025-06-17) and include EU standard contractual clauses (SCC, Module Two/Three) with a UK/Swiss addendum (anthropic.com/legal/data-processing-addendum); Anthropic's list of sub-processors: anthropic.com/subprocessors. [to be finalized by the lawyer: the legal basis of the transfer]
  • [Own-purpose data - controller side] Barion Payment Zrt. — handles the payment; an independent data controller. Payment is not switched on until the paid service launches.
  • [Own-purpose data - controller side] Cloudflare, Inc. (USA) — data processor: it performs the Turnstile bot-protection check (IP address, browser characteristics). According to Cloudflare's own Turnstile notice, it processes the signals on our behalf, and also uses them as an independent data controller to improve bot detection (cloudflare.com/turnstile-privacy-policy). [The legal basis of the transfer — to be filled in based on legal review.]
  • [Own-purpose data - controller side] Google Ireland Limited — an independent data controller in its own sign-in process (Sign in with Google); see policies.google.com/privacy.
  • [Own-purpose data - controller side] Számlázz.hu (KBOSS.hu Kft.) — issuing and sending electronic invoices. [The role as data processor/controller and the company data — to be checked.]
  • [Content - sub-processor] Hosting provider — Hetzner Online GmbH (Industriestr. 25, 91710 Gunzenhausen, Germany) — the server is located in the European Union (Germany). The database (accounts, summaries) and the server-side backups are here. We concluded Hetzner's data processing agreement (GDPR Art. 28, version 1.2) on October 27, 2026.
  • [Own-purpose data - controller side] E-mail sending provider — Brevo (SAS, formerly known as Sendinblue, a French company; RCS Paris 498 019 298; registered office: [to be checked — Brevo gives two addresses: 9-17 rue Salneuve, 75017 Paris, and 106 boulevard Haussmann, 75008 Paris]), for password reset and account-related e-mails, and for notices about changes (source: brevo.com/legal/privacypolicy). E-mail sending works only if the Provider has switched it on.
  • [Content - sub-processor] Microsoft (OneDrive) (Microsoft Ireland Operations Limited [to be checked: company data]) — storage provider for the encrypted backup: the copy is in the Provider's personal OneDrive storage, as a password-encrypted file. The whole database backup — so the Content too — reaches it in this encrypted file. Microsoft cannot access the content of the file without the key; it does not receive the password (see section 5). [lawyer: the legal classification of Microsoft (processor or not) — to be filled in based on legal review.] [to be finalized by the lawyer: listing it as a sub-processor] Personal (consumer) Microsoft account; the contractual chain is not verified. [open: a suitable contract or stopping the synchronization]

We do not sell your data, do not share it for advertising purposes, and do not use it for profiling.

4. Why we process this data (legal basis)

  • [Own-purpose data - as controller] We process the data concerning you that is necessary to create the account, to sign in, to manage the platforms and devices you have connected, and to perform the requested service, on the basis of performance of the contract. The processing of conversation content relating to other persons, of our own security purposes and of mandatory accounting data is to be assessed separately under the points below.
  • [Content - as processor] We process the content of the conversations — including the data of others, among them children — at your request, on your behalf and instructions, to make the summary meant for you, as a data processor; we do not use it for our own purposes. You are responsible for having a legal basis for uploading the content; in private, household use the exemption under Article 2(2)(c) GDPR may apply to you. [to be finalized by the lawyer: the legal basis (legitimate interest / the applicability of the household exemption on the user's side), the special-category data (e.g. data pointing to health) and the processing of children's data, and the balancing of interests] See section 7.
  • [Own-purpose data - as controller] Of the data received during Google sign-in, we process the e-mail address and the Google ID to create the account and to sign in, on the basis of performance of the contract; the temporary receipt of the name and the fact that it is not stored are described in section 2.
  • [Own-purpose data - as controller] We carry out the bot-protection check based on our legitimate interest in protecting the service.
  • [Own-purpose data - as controller] We process invoicing data based on a legal obligation (accounting and tax law requirements).
  • [Own-purpose data - as controller] We process feedback based on our legitimate interest (improvement, bug fixing).
  • [Own-purpose data - as controller] We process the usage log and the abuse-protection data based on our legitimate interest in running the service securely and sustainably.

5. How long we keep it

By role: for the summaries, the raw text and the groups (Content - here we are data processors) the periods set on your instructions apply; for the account, the session, the logs, the feedback and the documents (own-purpose data - here we are data controllers) our own retention periods apply.

  • Data belonging to an account (account, platforms, devices): until your account is deleted.
  • Summaries: we keep them for 90 days, then they are deleted automatically. If a summary contains a later date or deadline, you can extend it with the "Keep" button: we keep it for 14 days after the latest date in it, but for at most 1 year counted from the creation of the summary. You can also delete any summary yourself, one by one, at any time.
  • Summaries made earlier without an account: only a registered user can make a summary. Summaries made earlier without an account can no longer be opened, and we delete them.
  • Raw message text: we do not store it permanently on our server; it is present in the server's memory while it is being processed, and the rules of this notice also apply to that handling. At the AI provider it may be kept for a limited time under its own terms, as described in section 3 (according to its public information, by default for about 30 days, with exceptions) [to be verified: the retention under the contract and API setting actually applied].
  • Session: 30 days (the cd_session cookie); password reset link: at most 60 minutes, single use.
  • After your account is deleted: we end access to the active account and the summaries, and we carry out the deletion of active data in line with the following backup paragraph and with the lawful retention exceptions. [The exact active deletion steps and the deadline are to be filled in after technical verification.] We end the renewal of the subscription; the payment access data, the mandatory documents and the data of any settlement in progress are to be kept separate. We delete the feedback linked to the account (e-mail address and text). This alone does not make the remaining log anonymous; the legal basis and identifiability of any further retention must be examined separately. Invoices and tax documents that must be kept by law are not deleted with the account: they are kept for the tax-law period applicable to them. [For each type of document, the exact legal basis, the period or how it is calculated, the starting point and any extending cause are to be given.] We keep consumer complaints and a copy of the substantive reply for three years.
  • Backup: the database is backed up daily; we keep a backup on the server for at most 30 days, after which it is deleted automatically. We also keep a copy of the backup outside the server, in the Provider's personal Microsoft OneDrive storage, in a password-encrypted (AES-256) zip file; only the Provider knows the password, and it is not put in the cloud. Together with the OneDrive recycle bin, the off-site copy can exist for at most about 44 days. Deleted accounts and summaries may therefore remain in the backups on the server for up to 30 days and in the off-site copy for up to about 44 days. If we restore from a backup, we apply the earlier deletions (one-by-one, expiry and account deletion) again. For this we keep the identifiers of deleted summaries (without content) in a deletion log for 45 days: this way a deletion can be re-applied until the oldest backup expires; after that the log entry is deleted too.

6. Your rights

You can delete your account and your summaries yourself, taking into account the active-deletion and backup rules in section 5; you can revoke the devices that have access to your account; you can disconnect your platforms. In addition, under the conditions set by law, you can request a copy of your data, rectification, restriction of processing, you can object to the processing, and you can request your data in a portable format at varga.fabian.gyorgy@gmail.com. We reply to your request without undue delay, as a rule within one month. If the deadline may lawfully be extended because of the complexity of the matter or the number of requests, we inform you of this, with reasons, within the first month. You can lodge a complaint with the National Authority for Data Protection and Freedom of Information (NAIH, naih.hu) or go to court.

For your own-purpose data (account, payment, logs, feedback) we are the data controller, and your request must be addressed to us. For the Content we act as a data processor: if your request belongs to the controller (the user), we forward it to them and, at their request, help to carry it out (see section 7 and the Data Processing Terms).

7. Data of others and children in conversations

Group conversations also contain messages and data of people who are not users of Chat Digest — typically other parents, teachers, coaches, and often information about children. We use these only to make the summary meant for you: we do not store the raw text, do not build a profile from it, and do not use it for any other purpose. The finished summary may, however, contain this data, and it remains until it is deleted automatically (section 5) or until you delete it or your account.

If you believe a summary contains data about you or your child, you may write to varga.fabian.gyorgy@gmail.com even if you have no Chat Digest account. Give only the least information needed to identify the content concerned; do not send a whole conversation or an identity document unnecessarily. Where justified, we may ask for proportionate further information to confirm identity or authority to represent, and we do not automatically release other people's messages. We will examine your request within the deadline in section 6, under the statutory conditions, and where justified we will rectify, delete or restrict the processing of the data concerned. You can also lodge a complaint with NAIH. We receive a request concerning the content confidentially; after the necessary identification and role check, we involve the principal user without undue delay and help with the lawful settlement. We decide on requests about our own data processing under GDPR Art. 12. Even in household use we provide a direct contractual complaint channel; this does not pass automatic GDPR obligations to the parent. The user is responsible for having a legal basis for uploading the content; in household use the exemption under Article 2(2)(c) GDPR may apply to the user. [to be finalized by the lawyer: the forwarding of requests, partial rectification/deletion, the actual process of verifying representation, the reference to the household exemption and the information under GDPR Art. 14] We recommend that you inform the other participants of the conversation in advance that you are having it summarized; they may also contact us directly at the e-mail address above with their questions. [to be finalized by the lawyer]

Only a person aged at least 18 may use the Service. Children's data may get into the summary through parents' conversations; the rules above apply to this.

8. Cookies and local storage

We use only technically necessary cookies, which do not need separate consent. We have no analytics or advertising cookies, and we do not use third-party tracking code. To improve the service we keep a daily total count of how many times the home page, the upload page, the pricing page and the registration page were loaded (excluding bots), how many people registered, and how many accounts made their first summary. For this we do not use a cookie, and we do not store the IP address, the browser data, the user's identifier or the parameters of the visited addresses — only the daily count.

Cookie / storagePurposeLifetime
cd_sessionthe session identifier that keeps you signed in30 days
cd_langremembering the chosen language (functional, necessary cookie): it keeps only the language choice (hu/en) and cannot identify you 6 months
State/nonce cookie of Google sign-in (and the terms-acceptance flag)protecting the "Sign in with Google" process10 minutes
Cloudflare Turnstilebot protection (on the registration page)[to be checked — see Cloudflare's cookie notice]
The extension's local storage (chrome.storage.local)the address of the Chat Digest server, the device's identifier and token, the account's e-mail addressThe local data are deleted if you disconnect, switch to another server or e-mail address, or the server reports the token as invalid (in that case the device identifier and the token are deleted); when the extension is removed, the browser deletes the whole local storage. Revoking the device means that the token is invalidated on the server side; immediate deletion of the local data on an offline device is a separate question: there the local data may remain until the next connection or until the extension is removed. [to be finalized by the lawyer]

9. Data security, access to summaries

  • The website is available over HTTPS. We store passwords in non-reversible form. The detailed security measures for processing the Content are described in section 6 of the Data Processing Terms.
  • No other user can open a summary belonging to your account, only you, when logged in — its link does not work for anyone else. For operating the Service the Provider has technical access to the database, but looks at the content of summaries only for troubleshooting or because of a legal obligation.
  • The browser extension does not ask for, see or store your Messenger/Facebook password; it has no access to social pages (see section 2). The extension keeps the access token belonging to your Chat Digest account in your browser's local storage.

10. Scope of use

Chat Digest is meant for private, household use: so that you, as a natural person, can review your conversations in order to manage your own and your family's affairs. It may not be used for institutional (e.g. nursery, school, association), employer, business or monitoring purposes; the details are in section 5 of the Terms. The organiser of a parents' association or a class parents' group is a private user if they ask for a summary for their own family's information; not if they keep records on behalf of the community or of an institution. [The wording is awaiting legal review.]

11. Contact

For questions or requests: varga.fabian.gyorgy@gmail.com

12. Changes

We may update this notice from time to time; we will inform you of material changes by a notice published on the website, and we will inform registered users by e-mail. Before any new purpose of processing we will inform you in advance, and amending this notice does not by itself give a right to use the content in a new way.

← Back

Add to home screen (iPhone, iPad)

  1. Open this page in Safari (in other browsers the menu below may differ).
  2. Tap the Share icon at the bottom (□↑).
  3. Scroll down: Add to Home Screen.
  4. Tap Add.

Our server does not store the original messages; the digest is made by an external AI provider. · Terms · Privacy notice · Data processing terms

Barion — accepted payment methods

Feedback

Describe the problem — don't paste conversations or other people's personal data here.