Privacy Policy
Effective 20 October 2026
This policy covers the aireader app for iPhone and iPad, the web app at https://aireader.news/app/, and the aireader.news service both connect to. All three are operated by Inal Kardanov, an individual in the United Arab Emirates ("we"). Questions: [email protected].
1. The short version
- The app is an RSS reader. Our server fetches the feeds you follow and prepares the articles; the app downloads them so you can read offline.
- To do that we keep an account for you, the list of feeds and folders you set up, and which articles you have read or saved.
- We do not show ads, do not track you across apps or websites, do not use analytics or advertising software, and do not sell or rent your data.
- You sign in with Apple; where the web app offers it, you can sign in with Google instead or as well. We never see your Apple or Google password. We do not ask Apple for your name or email address. If you sign in with Google, Google sends us your name, email address and the address of your profile picture with every Google sign-in, and we keep them with your account (section 2).
- You can export your subscriptions at any time, and delete your account and everything in it from inside the app or the web app.
- If you choose to connect your ChatGPT account for summaries, our server stores your OpenAI sign-in tokens in encrypted form and sends article text to OpenAI under your account. This is off unless you turn it on.
- The web app is the same reader in a browser. It talks to the same server, keeps your articles in the browser's own storage on your computer, sets no cookies and loads nothing from anyone but us and, for the sign-in, Apple. Signing in with Google takes you to Google's own page and back.
2. What we store on the server, and why
| Data | Details | Why |
|---|---|---|
| Your account | Our internal number for the account, the date it was created, and the sign-in methods connected to it (the next two rows), each with the date it was added | To keep your account |
| Sign in with Apple | The identifier Apple gives us for your Apple Account (specific to this app's developer). We do not ask Apple for your name or email address. If the sign-in data Apple sends contains an email address all the same, as it can for an account created with a pre-release version of the app, we store that address and the name given then, and Settings → Account shows the address | To recognise you when you sign in with Apple and show the account in Settings |
| Sign in with Google (web app) | Only if you sign in with Google or add it to your account: the identifier Google gives us for your Google Account, and what Google sends with each Google sign-in: your email address and whether Google has verified it, your name, first name and last name, the web address of your profile picture and, if your Google Account belongs to a Google Workspace or Cloud organisation, its domain; and when Google was added and last used to sign in. At each Google sign-in these details are replaced and no earlier version is kept on the server; copies in backups go on the schedule in section 6. We never download or show the picture. We do not keep the access token Google issues. The iPhone and iPad app never receives any of these details | To recognise you when you sign in with Google, to show which Google Account is connected, and to know where we may answer you (section 8) |
| Sign-in sessions | For each device or browser you sign in on: a random identifier the app creates for that installation (the web app, for that browser profile), a device label (the text "aireader on iPhone" or "aireader on iPad"; from the web app the browser and system, for example "Safari on Mac" or "Chrome on Windows"), when you signed in, when the session was last used (updated at most once an hour), when it expires and, in the web app, when you last confirmed it's you before changing a sign-in method (cleared when the change is made). We store a one-way hash of the session token, not the token | To keep you signed in and to let you sign out of one device or all of them |
| Folders and subscriptions | Folder names and their order; which feeds you follow, the folder each one is in, any title you gave it, and when you added it | This is your reading list. It syncs to your devices |
| Reading state | For each article delivered to you: whether it is read, whether it is in Read later, and when each of those changed | So that read and saved articles are the same on every device |
| Settings that sync | Whether summaries are automatic, the summary language, the daily summary limit, and the language code of your device (for example "en"), which the app sends on its own so that summaries can be written in your language | To apply your choices on the server |
| Sync bookkeeping | A record of what changed and when, so each device can fetch only what is new; markers for things you deleted; the identifiers and results of changes your device sent; one-time request keys with a fingerprint of the request and the stored reply | So that changes made offline are applied exactly once and nothing is lost |
| Summaries | The summaries generated for you, each with the model and language used, and a counter of how many were generated in the last day | To show them on your devices and to enforce your daily limit |
| ChatGPT connection | See section 4 | To generate summaries with your own account |
| Technical logs | For each request: time, address and method, response status and duration, a request identifier and the IP address it came from. Background jobs may log our internal number for your account. Sign-in tokens, one-time sign-in codes and authorization headers are removed before anything is written, and the details Apple and Google send about you are not written to the logs | To find faults and stop abuse |
Sign-in methods. An account can have one Apple and one Google sign-in. Where the web app offers Sign in with Google, you can add the second one under Settings → Account → Sign-in methods; either then opens the same account. You can remove either one there, but not the last. Removing a sign-in method deletes its identifier and every detail that came with it from our server at once; copies in backups go on the schedule in section 6, and if a backup is ever restored, the removal is repeated from a record we keep for that purpose.
Information from Google is used only to sign you in, to show which Google Account is connected and to know where we may answer you when you write to us. We do not sell it, use it for advertising, share it with anyone, send it to OpenAI or use it to train AI models. Our use of information received from Google follows the Google API Services User Data Policy, including its Limited Use requirements.
Request limits (for example, how many sign-in attempts an address may make per minute) are counted by IP address in the server's memory and are not written to disk.
Articles are not personal data. Feeds, articles, the extracted text and the images are stored once on the server and shared by everyone who follows the same feed. They say nothing about you. The fact that it is you who follows a feed is part of your subscriptions, above.
Feed addresses. When you add a feed, its address is stored on the server, in a list shared by all accounts. If a feed address contains a private key or token of yours, be aware that it is stored as you entered it. The address stays in that list after you unsubscribe or delete your account; it is then no longer linked to you.
OPML files you import are read to create your subscriptions and are not kept. The list of feeds found in the file is stored with the request key for 30 days.
Discover. The Discover catalog is a list that ships inside the app and is refreshed from our server as one file; the web app fetches the same file when you open Discover. Which sources you look at is not sent anywhere. Adding a source from it is the same as adding its feed by hand.
3. What we do not collect
Your name or email address from Apple, except as described in section 2. Location. Contacts, photos, microphone or camera. Advertising identifiers. Payment details. What you type into search: search runs on your device and never leaves it. Crash reports or usage analytics: the app contains no third-party code at all, and the web app contains none of ours or anyone else's beyond one script from Apple that performs the sign-in (section 7). Apple may show us crash and usage statistics from people who agreed on their device to share them with app developers; they do not identify you.
Our server sets no cookies, and the pages of this website and the web app set none. The web app keeps its sign-in in the browser's own storage, not in a cookie, so it is never sent anywhere by itself.
4. Summaries with your ChatGPT account
This feature is optional, and nothing is sent until you connect an account. The app shows what is described here before you do.
- Connecting. The app shows a one-time code. You enter it on OpenAI's own sign-in page. We never see your OpenAI password.
- What we then store. The access, refresh and identity tokens OpenAI issues, encrypted with AES-256-GCM under a key that exists only on our server; your ChatGPT account's email address, ID and plan, as stated in those tokens; the model you chose and the list of models available to your account; when the connection was made and last refreshed; and, if OpenAI reports a usage limit or an error, until when and which. The tokens are kept on our server and never on your device.
- What is sent to OpenAI. For each summary, our server sends the article's title, source, author, date, address and text (long articles are shortened to about 22,000 characters), with instructions to summarize it. The request is made with your tokens, so it is sent under your account and counts toward your ChatGPT usage limits. We ask OpenAI not to store the exchange (
store: false). We do not send your Apple or Google identifier, the details Google sends us, your subscription list or your reading history. - Not an OpenAI product. OpenAI does not endorse this feature. It signs in the way OpenAI's Codex tool does, so OpenAI may change or block it at any time, and your use of your ChatGPT account remains subject to OpenAI's terms.
- OpenAI's own handling of those requests is governed by your agreement with OpenAI and its privacy policy, not by this one.
- Disconnecting (Settings → AI Summaries → Disconnect) deletes the stored tokens and those account details from our server at once. It does not sign that session out at OpenAI; to do that, use "Log out of all devices" in your ChatGPT settings.
- Sign-in attempts that were started and not finished are deleted after one day.
5. What is stored on your device
- A database with your account's identifier and the address of the server it belongs to, your folders, subscriptions, articles, reading state and summaries, and the downloaded article text and images. If Apple gave us an email address or a name (section 2), the database holds a copy of them. The app never receives the details Google sends. Downloads are excluded from iCloud and computer backups.
- In the iOS Keychain, on this device only: your session token, the installation identifier and the Apple account identifier.
- Your settings (reading, appearance, offline and text options).
Signing out erases the database, the downloads, the session token and the Apple account identifier from the device. The installation identifier stays in the Keychain so that the device is recognised if you sign in again. It is a random value that means nothing outside this app, and iOS may keep it even after the app is deleted. Deleting your account from a device removes the database, the downloads and all three Keychain items from that device, the installation identifier included. Your reading, appearance and offline settings stay; they say nothing about the account.
The app reads the clipboard only when you tap "Paste link from clipboard", and writes to it only when you tap a Copy button or, while connecting ChatGPT, "Open sign-in page", which copies the one-time code so that you can paste it on OpenAI's page.
The app can be pointed at a different server under "Advanced" on the sign-in screen. If you do that, this policy does not apply to that server; whoever runs it is responsible for your data there. iOS asks for local-network permission only if the address you enter is on your local network.
In the web app
The web app at https://aireader.news/app/ keeps everything in the storage your browser sets aside for that site, on the computer or device the browser runs on, and nowhere else:
- One database with the same contents as the app's: your account's identifier and the address of the server, your folders, subscriptions, articles, reading state and summaries, the downloaded article text and images, the changes you made that have not reached the server yet, the Discover catalog it last fetched, your settings, your session token and a random installation identifier. If an email address or a name is stored with your account (section 2), from Apple or from Google, the database holds a copy of them, with the sign-in methods the account has and when each was added.
- A copy of the web app's own files (its page, scripts, styles and icons), kept by the browser so that the web app opens without a connection. It holds none of your data.
- While you sign in with Google, the browser tab keeps in its session storage a random value that ties the sign-in to that tab, the sign-in's reference, the screen to return to and, when you are adding Apple, Apple's one-time sign-in answer. They are removed when Google sends you back, or when the tab is closed.
Nothing is kept in a cookie. Apart from signing in with Google and changing sign-in methods, which only the web app offers, no information about you is sent to our server that the app would not send. The web app always connects to aireader.news; it cannot be pointed at another server.
Signing out erases the database, the downloads and the session token from the browser. The installation identifier stays, so that the browser is recognised if you sign in again; it is a random value that means nothing outside the web app. Deleting your account from the web app removes the database, the downloads and the installation identifier as well. The browser's own controls remove the same things: clearing the site's data for aireader.news, or "Clear history and website data", deletes everything the web app stored, and you are signed out of it. Safari also deletes a site's stored data by itself after seven days of Safari use without a visit to the site; the web app says so in Settings → Offline on Safari. The web app asks the browser once, after you sign in, to keep its storage; a browser that does not promise to may remove the downloaded articles when space runs low, and Settings → Offline then says so.
The web app writes to the clipboard only when you tap "Copy code" while connecting ChatGPT, or when you share an article in a browser that has no share sheet and the link is copied instead. It never reads the clipboard. Sharing hands the article's title and address to your device's share sheet when the browser has one.
6. How long we keep it
| Data | Kept for |
|---|---|
| Account | Until you delete it. An account is not removed for being unused. The details of a Google sign-in: until you remove Google from the account or delete the account; they are replaced at each Google sign-in |
| Sign-in sessions | 180 days after last use. Expired sessions are removed at the next nightly clean-up; a session you signed out of is removed 30 days later |
| Folders and subscriptions | Until you change or remove them, or delete the account. When you unsubscribe from a feed, its articles are removed from your account at once, except those in Read later; the subscription record goes once no saved article from it remains |
| Reading state of articles you have read | Deleted 60 days after the article reached your account, unless it is in Read later |
| Unread articles | Marked read automatically once the article is more than 30 days old, or when one feed has more than 500 unread, then deleted as above |
| Read later | Until you remove the article from Read later or delete the account |
| Markers for deleted items | 30 days |
| Records of applied changes and one-time request keys | 30 days |
| Summaries | As long as the article's text is kept on our server, which can be longer than the article stays in your list, or until you delete the account. The daily counter: 2 days |
| ChatGPT connection | Until you disconnect or delete the account. Unfinished sign-in attempts: 1 day |
| One-time sign-in codes used with Apple or Google, and Google sign-ins that were started and not finished | 10 minutes; every value that could be used to sign in is stored only as a hash; deleted at the latest the day after they expire |
| Technical logs | Kept in files of limited size and overwritten by newer entries; they are not archived. Entries about a deleted account (IP address, internal account number) are not edited; they go when they are overwritten |
| Shared articles, text and images | An article nobody's account refers to any more is removed about 90 days after it was fetched; its text and images follow within about a week |
| Shared feed addresses | Kept. They are not linked to an account once nobody follows them |
| Backups | The whole database is backed up daily on the same server. Encrypted copies of the backups are also kept off the server, on storage we hold ourselves in the United Arab Emirates, so that the data survives the loss of the server; no outside storage provider is involved. A backup, and every copy of it, is deleted after at most 35 days, so data you delete leaves the last backup within 35 days. If the database is damaged, a copy of it may be set aside for repair for at most 30 days |
| Record of deleted accounts | 180 days. It holds only our internal number of the account and the time it was deleted. If a backup ever has to be restored, the deletions made after that backup are repeated from this record before the service comes back |
| Record of removed sign-in methods | 180 days. It holds only our internal number of the account, which method was removed and when. If a backup ever has to be restored, the removals made after that backup are repeated from this record before the service comes back, except one that would leave an account with no way to sign in |
A deletion or a removal of a sign-in method made after our last off-site copy of those records can be lost together with the server, and would then have to be made again.
7. Who else is involved
We share personal data with no one for their own purposes. These parties take part in delivering the service:
- Apple — Sign in with Apple. Apple tells us the account identifier; we do not ask it for your name or email address. Apple knows that you use the app. Our server also downloads Apple's public signing keys to check each sign-in, and when you delete your account from the app it asks Apple to end the app's access to your Apple Account (section 8). In the web app the sign-in is Apple's own script, loaded from Apple's servers (appleid.cdn-apple.com) whenever the web app opens with a connection, and Apple's sign-in window; Apple therefore knows that your browser uses the web app. It is the only code in the web app that is not ours.
- Google — only if you sign in with Google in the web app. The web app sends your browser to Google's sign-in page and Google sends it back with a one-time code; no Google code runs in the web app. Our server exchanges that code with Google (oauth2.googleapis.com) for the details listed in section 2 and downloads Google's public signing keys to check them. Google therefore learns that you use aireader when you sign in with Google. Google's privacy policy applies to Google's handling.
- Cloudflare — network transport. All traffic between the app, the web app, this website and our server passes through Cloudflare's network, which protects the server and hides its location. Cloudflare therefore handles your IP address and the encrypted requests, and decrypts them at its edge before passing them to us through an encrypted tunnel. Caching is switched off for the service. This website, aireader.news, is served by our server too, through the same tunnel; its pages are plain files that contain no scripts and no trackers, and the web app's files under /app/ are our own program. Cloudflare's privacy policy applies to its handling.
- OpenAI — only if you connect your ChatGPT account (section 4).
- Publishers — the sites whose feeds you follow. Our server, not your device, fetches their feeds, article pages and images, so a publisher sees a request from our server and learns nothing about you. The exceptions are the things you open yourself: "Open original", "Open in Safari…" and links inside an article load the publisher's page directly from your device, as any browser would. "Open original" keeps no cookies between uses.
No advertising networks. No analytics providers. No data brokers. The article text is extracted by software running on our own server, not by an outside service.
We would disclose data if a law that applies to us required it, and would tell you unless the law forbade that.
8. Your choices
- Export. Settings → OPML → Export OPML in the app produces a standard file with all your subscriptions and folders. It is created on the device and works offline. The web app has no export.
- Remove things one at a time. Unsubscribe from a feed, delete a folder, or remove an article from Read later. Disconnect ChatGPT (Settings → AI Summaries in the app, Settings → ChatGPT in the web app). Remove downloaded copies from the device (Settings → Offline → Delete downloaded articles). Sign out of this device or browser, or of all devices (Settings → Account).
- Delete the account. Settings → Account → "Delete account…" in the app or in the web app deletes your account at once. It removes the account, its sign-in methods, your sessions, folders, subscriptions, reading state, summaries, any ChatGPT connection and the sync records from our server, and everything the app or the web app stored on that device or in that browser. It cannot be undone. If you no longer have the app, install it, sign in with the same Apple Account and delete the account there, or sign in to the web app and delete it from Settings → Account. Copies in backups go on the schedule in section 6.
- What stays after deleting. Articles, images and feed addresses are shared between accounts and stay; they are no longer linked to you. Log entries are not edited (section 6). A summary that was still waiting to be generated at the moment of deletion leaves a queued request that holds our internal number of the account and of the article; it produces nothing and is discarded when its turn comes.
- Sign in with Apple. When you delete your account from the app, we ask Apple to end the app's access to your Apple Account. If that cannot be done, the app says so and gives the one step that is left: in the iOS Settings app, tap your name, then Sign in with Apple, choose this app (it may be listed as "aireader" or "aireader: Offline RSS") and tap Delete. A deletion made from the web app cannot ask Apple at all, because the web sign-in gives our server nothing it could use for that; the web app shows the same step instead.
- Sign in with Google. Our server holds nothing it could use to end aireader's access to your Google Account. If you remove Google or delete your account in the web app, it shows the step that is left: open https://myaccount.google.com/linkedapps, choose aireader, then Stop using Sign in with Google.
- ChatGPT. Deleting the account deletes your OpenAI tokens, and the ChatGPT account's email address, ID and plan kept with them, from our server. OpenAI is not told. To end that sign-in at OpenAI as well, use "Log out of all devices" in your ChatGPT settings.
- Ask us. You may ask what we hold about you, ask for a copy, or ask us to correct it. Write to [email protected]. We will first ask you for details that let us identify your account; an email address alone does not. If your account uses Google with an address ending in @gmail.com, or with a Google Workspace address that Google has verified, we may write to that address, and we act on your request only after you have answered that message there. To delete your account, use Settings → Account in the app or the web app.
If you are in the European Economic Area, the United Kingdom or another place with similar law, you also have the right to object to or restrict processing and to complain to your data protection authority. We process account, subscription and reading data because it is necessary to provide the service you asked for; logs because we have a legitimate interest in keeping the service secure and working; the ChatGPT connection because you asked for it and can withdraw it at any time by disconnecting; and the Google identifier and the details Google sends with it because keeping them is part of how Sign in with Google works in aireader, as this policy says before you choose it (to provide the service you asked for). If you do not want them kept, sign in with Apple instead, remove Google if your account has another sign-in method, or delete your account.
9. Where it is kept, and how it is protected
The service runs on a single server operated by us in the United Arab Emirates. It is not rented cloud hosting. If you use the app from another country, your data is transferred to and stored in the United Arab Emirates. The encrypted backup copies kept off the server (section 6) are in the United Arab Emirates as well.
- Traffic is encrypted with TLS. The server accepts connections only through Cloudflare's tunnel and exposes no open port to the internet.
- There are no passwords to steal: sign-in is Apple's or Google's, and each sign-in is checked against that company's signing keys. Adding or removing a sign-in method needs a sign-in made just before, with one of the account's own methods.
- Session tokens are stored only as hashes. OpenAI tokens are encrypted at rest and tied to your account, so a copy placed in another account cannot be decrypted.
- We can technically read the database. We look at personal data only to run, secure and repair the service or to answer a request from you.
This is a small service run by one person. No system is perfectly secure. If we learn of a breach that affects your data, we will tell you and the authorities where the law requires it.
10. Children
The service is not directed at children. You must be at least 16 years old to use it. If you believe a child has created an account, write to us and we will delete it.
11. Changes
When this policy changes, the new version is published here with a new date. Earlier versions are available on request. We will not start using your data in a new way without publishing the change first.
12. Contact
Inal Kardanov