Privacy Policy of POLLAR P.S.A.

At Pollar, we respect your privacy. We want you to know how we collect your data, why we need it, and what rights you have. This Policy explains how we use your data and how you can exercise control over it.

This Policy describes the rules for processing data on our website https://pollar.news (the “Website”) and in our web, mobile, watch and TV applications, including Pollar News, Pollar Radar, Pollar Weather and Pollar Sports (the “Applications”).

We publish this Policy in Polish, English, German and French. If the versions differ, the Polish version prevails.

1. Who are we?

The controller of your personal data is Pollar P.S.A., with its registered office in Krakow at ul. Piastowska 46/12, 30-067 Krakow, registered in the Register of Entrepreneurs of the National Court Register maintained by the District Court for Krakow-Srodmiescie in Krakow, 11th Commercial Division of the National Court Register, under KRS number 0001194689, with NIP (tax identification number): 6772530681 (“Pollar”, “we”, “us”).

We are the creators of Pollar, a news aggregator that uses artificial intelligence to group and summarize press articles. In connection with this, we process the data of individuals who:

  • visit our Website,
  • use our Applications,
  • receive our emails, have subscribed to our newsletter or are on our waiting list (waitlist),
  • write to us or send us feedback,
  • visit our social media profiles (e.g., Instagram).

For matters related to the processing of personal data, you may contact us at the following email address: contact@pollar.news, with the subject line “GDPR”.

We have not appointed a Data Protection Officer. Write to the address above and the people responsible for data protection at Pollar will answer.

2. Where do we obtain your data?

As a general rule, the data we process is provided to us directly by you. This occurs when:

  • you visit our Website – in which case cookies or server logs are stored,
  • you use our Applications – you give us data when you create an account, set your preferences, follow people and topics, use the features described in this Policy, and when your device registers for notifications,
  • you subscribe to our newsletter or join the waiting list – you provide us with your email address so that we can send you marketing information, including information about the launch of our services and their updates,
  • you visit our social media profiles – you leave a trace in the form of a comment, like, or message that is visible to us and to the operator of the given platform.
  • you write to us or use the feedback form – you give us your name, email address and the content of your message, and, unless you untick “Attach technical details”, the page address, browser, window size, base text size and theme,
  • you buy a membership – you pay through Stripe, Apple or Google Play, and they tell us the status of your purchase. We never see your full card number.

Data that we receive indirectly is data that we obtain when:

  • you are the author of articles we analyse – data such as your name and the content of the article reach us automatically from the news data provider NewsAPI.ai (Event Registry d.o.o., Ljubljana, Slovenia), which collects publicly available articles from news sites.
  • you sign in with Apple or Google – these services pass us your email address and, if your account has them, your name and profile picture,
  • you appear in the news – public figures and other people named in published articles appear in our news material and on the people and topic pages we build from it. We take descriptions of people from Wikidata and photos from Wikimedia Commons and from the official photo services of public institutions. The section “People in our material” has the details.
  • your device connects to us – we derive an approximate country from your IP address, for example to choose which regional alerts you receive.

3. Purposes and legal bases for processing your personal data

Depending on how you use our Website and Applications, we may process your data for various purposes. Below you will find the most important ones:

You visit our Website:

  • to give you access to aggregated news and the features you use (Art. 6(1)(b) GDPR),
  • to ensure its proper functioning and security (Art. 6(1)(f) GDPR),
  • to analyze visit statistics and improve content (Art. 6(1)(f) GDPR).

Thanks to this data, the website runs smoothly and we can develop it in accordance with your needs.

You use our Applications:

  • to provide you with a service, i.e., to enable you to create an account and use the application’s features (Art. 6(1)(b) GDPR),
  • to personalise your feed and notifications from your settings, follows and, if you leave it on, your reading history (Art. 6(1)(f) GDPR, our legitimate interest in showing you the news that matters to you; see section 8),
  • to run the AI features you turn on: interpreting the preferences you type, Pollar Chat and search personalisation (Art. 6(1)(a) GDPR, your consent, which you can withdraw in Settings),
  • to sell and manage your membership (Art. 6(1)(b) GDPR) and to meet our accounting and tax obligations (Art. 6(1)(c) GDPR),
  • to show local news and weather for your location when you allow access to it (Art. 6(1)(a) GDPR, the permission you give in your device or browser).
  • to ensure the security of the service and protect it against abuse (Art. 6(1)(f) GDPR),
  • to analyze how you use the application in order to improve and develop it (Art. 6(1)(f) GDPR),
  • in the event of a potential dispute, for the purpose of defending or pursuing claims (Art. 6(1)(f) GDPR).

Without this data, we would not be able to provide you with access to your account or guarantee the security of our service.

You receive our emails or subscribe to the newsletter:

  • to send you sign-in codes, service messages, the daily Brief email you ordered and, if you are a member, member letters (Art. 6(1)(b) GDPR),
  • to send you marketing and promotional information, as well as to inform you about the launch of our services and their important updates (Art. 6(1)(a) GDPR in conjunction with the provisions of the Electronic Communications Act and the Act on Providing Services by Electronic Means).

You'll stay up to date with Pollar.

You write to us or send feedback:

  • to answer you and improve Pollar based on what you tell us (Art. 6(1)(f) GDPR).

You visit our social media profiles:

  • to conduct conversations, respond to comments and messages (Art. 6(1)(f) GDPR),
  • to present our offer and build our brand (Art. 6(1)(f) GDPR).

We use it to respond to your comments and messages.

You author articles we analyse or are named in the news:

  • to process and aggregate publicly available news content, which is the foundation of our service (Art. 6(1)(f) GDPR),
  • to identify you as the author of a cited article excerpt, fulfilling informational purposes and respecting copyright (Art. 6(1)(f) GDPR).
  • to report the news accurately, including who said or did what, as part of our journalistic activity (Art. 6(1)(f) GDPR).

4. What data do we process?

The scope of data we collect from you depends on how you use our Website and services:

If you visit our Website:

  • technical data: your IP address, the type of browser and device, language and an approximate country derived from your IP address;
  • if you enable browser notifications: a web push subscription issued by your browser, used to deliver notifications.
  • if you allow location in the browser: your location, used to find news near you. We do not save it in our databases; requests that contain it can appear in our technical logs, which are deleted after 7 days.

If you use our Applications:

  • if you have an account: email address; your name and profile picture if you sign in with Apple or Google and those accounts provide them; your sign-in sessions with a shortened IP address and the type of browser or device. Pollar accounts have no password;
  • in all cases: a pseudonymous device identifier generated by the app, device model, operating system, app version, language, time zone and an approximate country derived from your IP address; anonymous crash and performance diagnostics; and product-analytics events from a fixed list (see section 12);
  • if you enable push notifications: a push token issued by Apple, Google or your browser, together with your language, app version, notification choices and, when you are signed in, your account;
  • what you set up and save: feed preferences, including any text you type and the rules our AI derives from it; the people, topics, threads and teams you follow; topics you watch in Pollar Radar; bookmarks; your reading history, including time spent and how far you scrolled, if you leave it on; your search history, if you turn it on; and the notifications in your inbox;
  • if you are a member: your plan, membership status and purchase history; and, only if you choose, your name and tier on the public supporters page and the questions and replies you post as a member, which other readers can see, as well as your votes;
  • if you allow location: in Pollar News for Android, your location to find news near you, which we do not save in our databases (requests that contain it can appear in our technical logs for up to 7 days); in Pollar Weather, coordinates rounded to about 1 km, kept with your device only while a place has rain alerts or a Live Activity running;
  • if you use Pollar Chat: the messages you type. We do not keep chat conversations on our servers; your history stays on your device. If you report an answer, the reported exchange is kept in our technical logs for up to 7 days.

If you subscribe to the newsletter or join the waiting list:

  • email address and your subscription choices.

If you write to us or send feedback:

  • your name, email address and the content of your message, and, if you send feedback while signed in, which account sent it.

If you visit our social media profiles:

  • data that is publicly visible as part of your activity (e.g., first and last name or username, comments, likes, message content, profile picture).

If you author articles we analyze:

  • your name, if it appears in a publicly available article, together with the content of that article; and for people named in the news: their name, the role in which they appear, their statements, the Wikidata description and, for some people, a photo (see the section “People in our material”).

5. How long do we retain your data?

We retain your data only for as long as it is truly necessary, in accordance with the purpose of processing and applicable law. In the event that:

You visit our Website:

  • our technical logs are deleted after 7 days; our application logs do not record your IP address or what you search for. Cloudflare, our content delivery network, processes your IP address to deliver the Website and protect it from attacks,
  • cookies are retained in accordance with the information provided in the cookie banner and your browser settings.

You use our Applications:

  • we keep your account data for as long as you have an account and delete it when you delete your account (see below),
  • sign-in sessions are stored only with the network part of your IP address and the type of browser or device, and are deleted the day after they expire; a session expires 30 days after you last used it and at the latest 90 days after you signed in,
  • anonymous crash and performance diagnostics and our technical logs are kept for up to 7 days; security audit logs, which contain your full IP address, for up to 90 days,
  • product-analytics events for up to 90 days, your reading history for 30 days, your search history (if you turned it on) for 90 days and your notification inbox for 90 days,
  • device records and the push token of a device that has not used Pollar for 400 days are deleted; Weather locations are deleted within a day after you turn off rain alerts for a place or its Live Activity ends.

You subscribe to the newsletter or join the waiting list:

  • we retain your data until you unsubscribe from the mailing list (by withdrawing your consent).

You write to us or send feedback:

  • we keep your messages for as long as we need them to deal with your matter, and afterwards only as long as claims related to it could arise.

You visit our social media profiles:

  • we retain your data for as long as you follow our profiles or engage in interactions (comments, messages),
  • your reactions (e.g., likes) may remain visible for longer, unless you remove them. We use this data to be able to respond to your comments or messages.

You author articles we analyze:

  • we keep the full text of the articles we fetch, which may include the names of their authors and of the people mentioned in them, for up to 30 days after fetching it (sports articles: up to 60 days); after that it is deleted automatically (backups that still contain it are overwritten within the following 30 days). Texts fetched before we introduced this limit on 13 October 2026 are being deleted gradually, every night; we will clear that backlog by 31 October 2026. After that we keep only what we need to cite the article as a source of published material and to recognise what it was about: its title, link, source name and publication date, a numerical fingerprint of the text (embedding) and the identifiers of the public figures and places the article concerns. Published material may also contain short quotes from sources with the name of the person quoted. We do not keep authors’ names as a separate field.
  • we do not collect articles from publishers who have reserved text and data mining for everyone in a machine-readable form (such as TDMRep), or who have told us they reserve it. Publishers can send such a reservation to contact@pollar.news.

Account deletion: where, how, and what stays

We follow Apple's App Store rule 5.1.1(v), Google Play's account deletion requirements and Art. 17 GDPR. You can delete your Pollar account inside the app or on the web. No phone call, no support ticket.

  • In the iOS and Android apps: Settings, then Privacy, then Delete account. The confirmation screen explains the 30-day grace period before irreversible deletion.
  • On the web: sign in at https://pollar.news, open Settings, and use the same Delete account flow.
  • During the 30-day grace period: your account is disabled and we stop sending you notifications, emails and Radar briefings. Sign in within the grace period to cancel deletion. After 30 days, the data listed below is irreversibly removed.

Removed at the end of the grace period:

  • your profile, e-mail and authentication data,
  • your feed preferences, free-text preference, the interpreted personalisation rules and your AI consent records,
  • your follows, Radar topics and briefings, bookmarks, reading and search history, notification inbox and notification preferences,
  • your active sessions, the devices only your account used (with their push tokens, Weather locations and Live Activities) and the product-analytics events of those devices,
  • your membership records with us, your customer record at Stripe and your Radar audio briefings; your name is removed from the supporters page.

Retained, with a documented purpose and a defined retention period:

  • Payment records: invoices and transaction records for payments on the Website are kept for 5 years, as required by Polish accounting and tax law, in Stripe on our behalf and in our accounting records. Purchases in the apps are sold and recorded by Apple and Google Play under their own obligations. We keep nothing else linked to you.
  • Security audit logs (sign-in events and the record of your deletion request): up to 90 days, for incident response and fraud investigation. After deletion these entries no longer contain your device, IP address or browser data, and the record of the deletion keeps only your bare account number, so that we can repeat the deletion if we ever restore a backup.
  • Backups: database backups are kept for up to 30 days and then overwritten. We use them only to recover from a failure, and if we restore one, we repeat every account deletion made since it was taken.
  • Anonymised content: questions you asked as a member that were answered publicly or printed in a member letter stay, signed “Deleted account” and no longer linked to you; everything else you posted as a member is deleted. Topics you watched in Radar stay as shared topic records with no link to you, because other readers may watch them too.

Once the grace period ends, the irreversible deletion runs automatically, within an hour. We do not keep a soft-deleted version of your account beyond that point.

Step-by-step instructions, and how to reach us if you can no longer sign in, live on a dedicated page: How to delete your Pollar account.

6. To whom may we disclose personal data?

We may share your data with trusted partners and providers who help us deliver our services. Depending on how you use Pollar’s services, your data may be shared with the following categories of recipients:

Everyone who visits the Website or uses the Applications:

  • hosting and infrastructure: Hetzner Online GmbH (servers in Germany), which hosts our servers, databases and backups, and Cloudflare, Inc. (USA), our content delivery network, which processes your IP address to deliver Pollar and protect it from attacks, and which also keeps a second, off-site backup of our databases and files in data centres in the European Union (Cloudflare R2),
  • email: Resend, Inc. (USA) sends our sign-in codes, service emails, the Brief email, member letters and newsletters from its EU region (Ireland); our own mailboxes, which receive what you write to us and send through the feedback form, are hosted by Migadu (Switzerland),
  • notifications: Apple (Apple Push Notification service), Google (Firebase Cloud Messaging) and your browser's push service deliver notifications to your device. A notification can contain the name of a topic or person you follow.

You sign in with Apple or Google:

  • Apple and Google confirm your identity and pass us your email address and, if your account has them, your name and profile picture.

You buy a membership:

  • Stripe (Stripe Payments Europe Ltd, Ireland, and Stripe, Inc., USA) processes payments on the Website and receives your email address and an account identifier; Apple (App Store) and Google (Google Play) process purchases in the apps under their own terms. We receive the status of your purchase, never your full card number.

You set feed preferences, follow topics in Radar or post as a member:

  • the free-text preferences you type in Settings, the preference labels we use to suggest people and topics to follow, the text of topics you follow in Radar together with the text we generate about them, and the questions and replies you post as a member are sent through Requesty (Requesty Ltd, 71-75 Shelton Street, London WC2H 9JQ, United Kingdom), an AI routing service that acts as our processor. We use its EU endpoint: Requesty processes these requests in Frankfurt (Germany) and passes them only to AI models hosted in EU regions, on Google Cloud Vertex AI, Microsoft Azure OpenAI (for example, the Sweden Central region) and AWS Bedrock (for example, Frankfurt), while the decision whether a news item matters for a topic you follow in Radar is made by Cloudflare’s Clef model (Cloudflare Workers AI), hosted in the EU. On our behalf, the models turn your text into feed rules and follow suggestions, check and describe the topics you follow in Radar, track news about them and translate members’ posts. We do not send your email address or other account identifiers with these requests. Requesty does not store the content of these requests or the models’ answers and does not use them to train models; it keeps only content-free billing records (token counts, model, time and cost) for 6 years to meet statutory accounting obligations. To match a Radar topic against articles, the topic text and a description generated from it are converted into a numeric representation by Google Vertex AI in the EU; that request passes through OpenRouter, a US-based routing service, without any account or device identifier, and is sent only to a Google endpoint that does not retain the data. When you type a topic to watch in Radar, we also look up the words you typed in Wikidata’s public search service, run by the Wikimedia Foundation (USA), to suggest matching people, places and organisations; our server sends only the typed text, with no account or device identifier.
  • Radar audio briefings, available to members, are voiced by ElevenLabs (ElevenLabs, Inc., USA). The script names the topics and people you follow, and a topic you added in your own words is read as you wrote it. We send ElevenLabs only the text of the script, without your email address or any account or device identifier.

You use maps, local news or Pollar Weather:

  • Pollar Weather gets forecasts from Open-Meteo (Switzerland) and air-quality data from public measuring networks; our server sends them coordinates rounded to about 1 km, never your IP address or any identifier. Maps load map tiles from OpenFreeMap, and the place search sends the name you type to the OpenStreetMap Foundation's Nominatim service (United Kingdom); both requests come from your browser, so these services see your IP address. Maps in the Android app are provided by Google Maps.

You listen to Pollar Sounds in a podcast app:

  • download links in our podcast feed pass through OP3 (Open Podcast Prefix Project, USA), an open service that counts downloads. It sees the IP address and user agent of your podcast app and publishes only aggregate statistics.

You subscribe to the newsletter or join the waiting list:

  • Resend, Inc., which sends our emails, as described above.

You visit our social media profiles:

  • your data is visible to the operators of those platforms and is processed in accordance with their terms of service and privacy policies.

You are the author of articles that we analyze:

  • the content of publicly available articles (including the names of their authors and of the people mentioned in them) is processed by our technology partners, the providers of artificial intelligence models and of gateways to them (among them Anthropic, Google, OpenAI, also through Microsoft Azure, Cloudflare, DeepSeek models run by cloud providers, OpenRouter and Requesty), which summarise, translate and categorise it on our behalf, and by ElevenLabs, which reads our material aloud. Photos of people, together with the story’s title and the person’s name, are assessed by an Anthropic AI model, with a Cloudflare model as fallback, both through OpenRouter (see the section “People in our material”).

We do not send your email address or other account identifiers to AI model providers. They receive only the content needed for the features described in this section, and we use them only on terms or settings under which they do not train their models on it.

For advertising measurement, where it is on (in Europe only with your consent), we share a limited set of conversion events and hashed (SHA-256) identifiers with Meta, as described in the Advertising and measurement section below. We never sell your personal data, and beyond that limited measurement we do not share it with advertisers or data brokers.

Some of our providers are outside the European Economic Area (EEA). In the United States: Cloudflare, Resend, Stripe, OpenRouter, OpenAI, Anthropic, ElevenLabs, Meta, Apple, Google and OP3. For transfers to the US we rely on the European Commission's adequacy decision for the EU-US Data Privacy Framework (Commission Implementing Decision (EU) 2023/1795 of 10 July 2023) where the provider is certified, and otherwise on the standard contractual clauses approved by the European Commission (Art. 46(2)(c) GDPR). Requesty Ltd and the OpenStreetMap Foundation are in the United Kingdom, and Open-Meteo and Migadu in Switzerland; the European Commission has found that both countries ensure an adequate level of protection (for the United Kingdom, Commission Implementing Decision (EU) 2021/1772 of 28 June 2021, extended by Commission Implementing Decision C(2025) 8771 of 19 December 2025 until 27 December 2031; for Switzerland, Commission Decision 2000/518/EC). Because we use Requesty's EU endpoint and only models hosted in the EU, the content we send to Requesty is processed in the EEA. Cohere, which helps rank search results, is in Canada; the Commission has found that Canada ensures an adequate level of protection for commercial organisations such as Cohere (Decision 2002/2/EC). Public news text that we summarise may also be processed by AI providers in other countries to which OpenRouter routes requests; it is material that has already been published and contains no data about our readers.

These features send what you type or follow to providers outside the EU: Pollar Chat (Anthropic, with OpenAI as the fallback, through OpenRouter, USA); search (your query is turned into a numeric representation by OpenAI and results are ranked by Cohere, through OpenRouter); and, in Radar, the topic matching request described above, which passes through OpenRouter. We attach no account or device identifier to these requests and tell OpenRouter to use only providers that do not train models on the content. Radar audio briefings, which name the topics you follow, are voiced by ElevenLabs in the USA, also without identifiers. Everything else you type into Pollar's AI features, including your feed preferences, Radar topics and members’ posts, is processed by AI models hosted in the EU. You can withdraw your consent to Pollar Chat at any time in Settings.

7. What rights do you have over your data?

You have full control over your data. Under the GDPR, you may:

  • access your data – request information on whether we process your data and obtain a copy thereof,
  • rectify your data – correct your data if it is outdated or inaccurate,
  • erase your data (“right to be forgotten”) – request the deletion of your data. The fastest path is the Delete account flow described in section 5. If you need a deletion outside that flow, contact contact@pollar.news; we respond without undue delay and at the latest within one month.
  • restrict processing – temporarily suspend the use of your data, e.g., while we verify its accuracy,
  • data portability – receive the data you gave us in a structured format and transfer it to another controller, where we process it on the basis of your consent or of our contract with you,
  • object – object to the processing of your data where it is processed on the basis of our legitimate interest, including profiling,
  • withdraw consent – withdraw your consent at any time if we rely on it for processing (e.g., the newsletter, AI features, advertising measurement). Withdrawal of consent does not affect the lawfulness of prior processing,
  • lodge a complaint with a supervisory authority – if you believe that we are violating data protection rules, you may lodge a complaint with the President of the Personal Data Protection Office (UODO) in Warsaw.

To exercise your rights, you may contact us at contact@pollar.news, with the subject line “GDPR”.

People in our material

We report on events, so our material contains data about the people involved: their name, role, statements, a description of what happened and, for some people, a photo. This section is for them. Pollar publishes the daily “Pollar.pl” (see Contact). Under Article 2(1) of the Polish Personal Data Protection Act, Articles 13–16 and 18–22 of the GDPR, among others, do not apply to editing, preparing, creating and publishing press material. We nevertheless explain how we process this data and handle the requests described below. We also run the people pages and the archive of mentions on the basis of our legitimate interest (Article 6(1)(f) GDPR) in telling readers who is in the news and keeping the archive available.

Where this data comes from: articles by other outlets, which we receive from NewsAPI.ai (see section 2); descriptions of people, their roles and photos from Wikidata and Wikimedia Commons; photos from the official photo services of public institutions (among them the European Commission, the European Parliament, the Chancellery of the Polish Sejm, the US government, the Office of the President of Ukraine and the Kremlin). To find a photo, we send the person’s name to these services’ search. From photo libraries (Pexels) we only choose photos without recognisable people.

Photos: next to a story we show a portrait only of a public figure (someone who holds office, works in a public profession or is widely known) and only when the story’s title is about them. We do not show portraits of victims, suspects, witnesses or other private individuals there. On a person’s page we show the photo that Wikidata lists as their image, with its author and licence. To crop a photo well, our server automatically detects where the faces are. It does not recognise who the person is, does not create face templates and does not compare faces, so we do not process biometric data within the meaning of Articles 4(14) and 9 GDPR. We store only the resulting crop and, for portraits of athletes, the position of the face in the photo.

AI checks of photos: before a photo is used, an Anthropic AI model (with a Cloudflare model as fallback), called through OpenRouter (USA), checks whether it shows the right person and is fit to publish, for example that it does not show a victim, a child or a private individual. It receives the crop, the story’s title and lead, the person’s name and the photo’s description at the source; the providers may not train models on this data. We keep the result with a short reason, which may include the person’s name, so that the same photo is not assessed again. We keep a photo for as long as it is part of published material or of a person’s page.

Your rights and how to use them: if you are shown in our material, you can ask us to remove or replace a photo, to remove or anonymise a person page or older mentions, and you can object on grounds relating to your particular situation. You can have an untrue or inaccurate report corrected under the Polish Press Law (Corrections and complaints). Write to contact@pollar.news with the subject “Person in a story”, and include the link and your reason. We reply without undue delay and within one month at the latest. In deciding, we consider whether you are a public figure, how long ago the material appeared and whether it still matters to the public. You also have the right to lodge a complaint with the President of the Polish Personal Data Protection Office.

8. Personalisation and profiling

If you have an account or use the apps, we arrange your feed and choose which notifications to send you using your settings, the people and topics you follow, the categories you chose and, if you leave it on, your reading history. This is profiling within the meaning of the GDPR. We use it only inside Pollar, to show you news that matters to you; we do not use it for advertising and we do not share it.

No decision with legal or similarly significant effects on you is made solely by automated means (Art. 22 GDPR). You can turn reading history off in Settings, change or remove your preferences and follows at any time, and object to this profiling by writing to us; your feed then relies on general popularity and the categories you choose.

9. Children

Pollar is not directed at children. To create an account, turn on AI features or consent to advertising measurement you must be at least 16 years old. If you believe a child under 16 has given us personal data, write to us and we will delete it.

10. Analytical tools

On our Website we use a minimal, self-hosted analytics tool to understand overall traffic and service quality. How we measure our advertising is covered separately in the Advertising and measurement section.

Umami Analytics

We use Umami, a cookieless analytics tool that we host ourselves. It records page views, the referring site, country, browser, operating system, device type, screen size and language, as well as interaction and performance events, which can include the identifier of the article you opened. When you are signed in, it also notes that the visit is signed in, whether you have a membership and, if so, which tier; it never receives your account, email address or any other identifier. It does not use cookies or localStorage and does not store your IP address; it groups visits with an identifier that changes every month, so it cannot follow you over a longer period or across sites. We do not load it at all when your browser sends a Global Privacy Control or Do Not Track signal. We process this data under our legitimate interest (Art. 6(1)(f) GDPR). The data stays on our servers; requests pass only through Cloudflare, our content delivery network. Learn more.

11. Advertising and measurement

Like any modern publisher, we run advertising and measure whether it works, so we can spend responsibly and improve it. For this we use Meta (Meta Platforms Ireland Ltd and Meta Platforms, Inc.). The purpose is limited to measuring and improving our advertising; we do not use it to profile what you read.

What we share: only conversions: when you join the waiting list and when you first pay for a membership (with its amount and currency). With each conversion our server sends Meta, through its Conversions API, the SHA-256 hash of your email address and, for a purchase, of your Pollar account identifier. Where measurement is on, the Meta Pixel also loads in your browser: Meta then receives your IP address, browser information and our domain, and for a conversion the address of that page and of the page you came from. We do not send page views, we never share the articles you read and we never share anything you write in Pollar Chat. We do not use X (Twitter) or any other advertising partner.

Your choice and legal basis: in the European Economic Area, the United Kingdom and Switzerland we measure advertising only with your consent (Art. 6(1)(a) GDPR). Elsewhere it is on by default and you can turn it off; we honour the Global Privacy Control and Do Not Track signals.

Joint controllership: for collecting data through the Meta Pixel and sending it to Meta, we and Meta Platforms Ireland Ltd are joint controllers (Art. 26 GDPR). Meta is responsible for how it processes the data after it receives it, as described in its own privacy policy, and you can exercise your rights with Meta or with us.

Transfer to the United States: Meta may process this data in the United States. Meta Platforms, Inc. is certified under the EU-US Data Privacy Framework, and we rely on the European Commission's standard contractual clauses as a fallback.

Retention: Meta retains this data under its own data policy. We keep our record of your consent for the life of your account, so we can show when and how you made your choice.

How to turn it off: use the consent banner, change your choice in Settings, then Privacy, or send a Global Privacy Control signal from your browser. Turning it off stops new events from being shared.

12. Diagnostics and product analytics in our Applications

To keep Pollar stable and to understand which parts of the product help you, our Applications send two kinds of measurement data to our own servers. We use no third-party analytics SDK. This data is stored on our infrastructure in the European Union (Hetzner, Germany, with a backup copy at Cloudflare R2 in the EU), is never sold, and is never shared with advertisers or data brokers.

Anonymous diagnostics (crashes and performance):

Our iOS app forwards Apple's device reports about crashes, hangs and resource use (CPU, memory, disk), which Apple provides only if you allow your iPhone to share analytics with app developers; our Android app sends a report when it crashes, with the error message shortened to 500 characters. We store these reports without any device or user identifier and without your IP address, and we do not try to link them to you. Because they are needed to keep the apps working, there is no separate switch for them in Pollar (Art. 6(1)(f) GDPR).

First-party product analytics (how the app is used):

We record a fixed list of interface events so we can see where the experience works and where it breaks. Typical events are: onboarding steps, paywall viewed, subscribe tapped, a search being run (we never store what you searched for), an article shared, a tab selected, a session starting, the result of a notification-permission prompt, a bookmark added, a thread followed, and an error being shown. Some events include the identifier of the story, film or episode involved and how much of it you watched or read.

These events never contain free text or message content. They are tied to a pseudonymous device identifier, not to your account, and we do not analyse them by account. We use the link between your devices and your account only to delete these events when you delete your account and to count sign-ups per platform. They are not passed to any third party. We process them under our legitimate interest in improving the product (Art. 6(1)(f) GDPR).

Your control: in the apps, open Settings and switch off “Help improve Pollar” (in Pollar Radar: “Share anonymous usage”). It is on by default. On the Website these events are not sent when your browser sends a Global Privacy Control or Do Not Track signal. This setting is separate from the Reading history control described in section 8, and from crash diagnostics, which stay on.

We keep diagnostics for up to 7 days and product-analytics events for up to 90 days (section 5).

Search personalization

Search personalisation is off by default. When you turn on the controls in Settings, then Privacy, Pollar saves your search history for 90 days, orders suggestions and results to match what you read, and learns from which results you open. You can turn any of these off at any time, and clearing your data removes the searches and clicks tied to your account.

Separately, we count popular searches. We record how often a query is searched, clicked or returns nothing; to count each searcher only once, we keep a salted hash of the device identifier with the query for 8 days. It is not linked to your account or IP address.

Pollar News MCP server

Pollar runs a public Model Context Protocol server at mcp.pollar.news so AI assistants can read our published news. It exposes only public editorial content and has no access to your account, preferences, reading history, or any personal data. It cannot change anything.

The server needs no sign-in, sets no cookies and builds no profile of you. To prevent abuse it keeps the requesting IP address in memory for rate limiting, for at most 2 minutes, and never writes it to disk. It logs which tool was called, without the IP address. As in Pollar’s search, the text of a search query is turned into a numerical representation by OpenAI and the results are ranked by Cohere, both through OpenRouter (USA), without the IP address or any identifier. We do not sell or share this data. Requests to it pass through Cloudflare, our content delivery network.

13. Social media

We process the personal data of users who follow our profiles on social media platforms such as Facebook, Instagram, YouTube and TikTok and interact with them. Meta also acts as our advertising-measurement partner, covered in the Advertising and measurement section above. Detailed information on how each platform processes data can be found at the following links:

14. Cookies

On our Website we use cookies and similar browser storage (localStorage). They keep the Website working and secure and remember your choices.

The cookies we set ourselves are listed in the table below. In localStorage we keep settings in your browser, for example your colour theme, your consent choices, dismissed banners, your recent searches and your Pollar Chat history. This data stays in your browser and is not sent to us or anyone else, except your consent choices, which we also record with your account when you are signed in.

Essential cookies are needed for the Website to work and cannot be switched off. Marketing cookies are set only as described below, and you can manage them in the cookie banner or in Settings.

Umami Analytics, our analytics tool, does not use cookies or localStorage.

The Meta Pixel sets marketing cookies to measure our advertising, as described in the Advertising and measurement section. When a conversion is recorded, your browser may also send Meta the cookies that facebook.com has set, for example if you are logged in to Facebook. In the European Economic Area, the United Kingdom and Switzerland this happens only with your consent; elsewhere it is on by default and you can turn it off in the cookie banner or in Settings, then Privacy.

  • pollar_device_id

    A pseudonymous identifier of your browser that keeps notifications, your notification settings and personalised highlights working, with or without an account.

    1 year · Essential

  • __Secure-better-auth.session_token

    Keeps you signed in.

    30 days from your last visit · Essential

  • __Secure-better-auth.state

    Protects sign-in with Apple or Google against forgery.

    5 minutes · Essential

  • pollar_signed_in

    Tells the Website before it loads whether you are signed in (1 or 0, no account data).

    1 year · Essential

  • pollar_region

    Your region for consent purposes (Europe, United States or elsewhere), derived from your IP address, so we apply the right consent rules.

    180 days · Essential

  • NEXT_LOCALE

    Your language.

    Until you close the browser · Essential

  • pollar_had_highlights

    Reserves space for your highlighted stories so the page does not jump while it loads.

    30 days · Functional

  • pollar_seen

    Counts on how many days this browser has visited Pollar (a number and a date, nothing else), so a first visit gets a short introduction and fewer prompts.

    1 year · Functional

  • _fbp

    Set by the Meta Pixel when a conversion is recorded, to measure our advertising.

    90 days · Marketing

  • _fbc

    Set by the Meta Pixel when you arrive from a Meta ad, to attribute the click.

    90 days · Marketing

You can clear cookies and localStorage at any time in your browser settings. We honour Global Privacy Control and Do Not Track signals.

15. Changes to the Privacy Policy

We want our data protection rules to be transparent and aligned with how you actually use the Website. Therefore, this Policy may be updated from time to time.

The most current version of this document will always be available on this page, together with the date from which it takes effect.

This version of the Policy applies from: 17 October 2026