Privacy policy

Last updated: 14 August 2026 · Applies to the SummaReader app for Linux and Android, and to these pages.

The short version

SummaReader collects nothing about you. There is no account, no analytics, no telemetry, no crash reporting and no advertising. Your reading library is stored on your own devices, and when it is synced between them it is encrypted so that the server carrying it cannot read it.

The app does talk to the internet, because a reader must. This document is mostly about who it talks to and what they can see.

What stays on your device

Everything, unless a section below says otherwise:

None of this is sent to us. We operate no service that receives it — including sync, which is described below and which we do not host.

What leaves your device, and who sees it

WhenWhere it goesWhat they can see
You follow a sourceThe publisher's own serverYour IP address, and that something fetched their feed
You follow a YouTube channelYouTube, and i.ytimg.com for thumbnailsYour IP address, and which channel or video was asked for
You follow a Mastodon accountThe server that account is onYour IP address, and which account was asked for
You open a Nostr linknjump.meYour IP address, and which note was asked for
You open an articleThat websiteThe same as any browser visiting it
You summarize on your deviceNowhereNothing leaves
You summarize using an endpoint you configuredThat providerThe text of the article being summarized
You download a modelHugging FaceYour IP address and which model you chose
You turn on syncThe sync server you runEncrypted data it cannot read — see below

Two of these are worth stating plainly.

Fetching is direct. The app talks to publishers itself. There is no proxy in between, so publishers see your IP address exactly as they would if you visited in a browser — and we see nothing at all.

Remote summarization sends the article. If you point the app at OpenAI, OpenRouter, Anthropic, your own server or anything else, the text being summarized is sent there and that provider's terms apply to it. This is off by default; summarizing on your device sends nothing anywhere. The app tells you which of the two it is using.

Sync

Sync is off until you turn it on, there is no address in the box until you put one there, and it works between your own devices.

We operate no sync service. There is no hosted option and no default server: the only sync server your app talks to is one you or somebody you trust runs. The software for it is published separately, under AGPL-3.0.

A sync server cannot read your library. Everything is encrypted on your device before it is sent, with a key that exists only on your devices and is never transmitted to the server. That is a property of the design rather than a promise about anybody's conduct, which is the point: it holds even if the server is hostile.

What a sync server does hold:

If you lose your recovery code, nobody can recover your data — not you and not us. That is what "we cannot read it" means in practice.

What we collect: nothing

No analytics. No usage statistics. No crash reports. No device identifiers. No advertising identifiers. No third-party SDKs that collect anything.

The Android app requests three permissions: internet access, permission to show notifications if you enable them, and the camera.

The camera is used for exactly one thing: scanning the pairing code shown by a device you already use, when you add a second device. The scanner reads one code and closes. Nothing is photographed, stored, or sent anywhere, and the same pairing code can be pasted as text instead if you would rather never grant it.

The local API

The app can serve a small HTTP API on your own machine — a door for a browser extension or a script to put something into your library, and a /metrics endpoint for a dashboard. It is off by default, there is no default token and no default port, and nothing listens until you switch it on.

Two things to know if you do. It requires a token you set, and it can be bound to every network interface rather than only this machine — which is useful and is also your decision to make, because behind that token is your library in plaintext.

Purchases

There are none. The app has no store code, no licence keys and no payments, so there is nothing here about billing. If that changes, this section is where it will say so.

How long things are kept

On your device, by default: article text for 14 days, and items for 180 days, after which the app tidies them away. Saved items are never removed automatically. You can change both, and you can delete everything at once from Settings.

On a sync server, until you delete it. There is a wipe operation that removes your synced copy; it is deliberately separate from switching servers and from anything else, so it cannot happen by accident.

Your rights

Under the GDPR you have rights of access, correction, erasure, portability and objection. They are unusually simple to exercise here, because we hold almost nothing:

Children

The app is not directed at children and collects nothing from anybody, including them.

These pages

There is no service here to sign in to — a product page and a user guide is all this is. They set no cookies, run no analytics and load nothing from any other host: no fonts, no scripts, no embedded anything. Whoever serves them records what any web server records in order to hand over a page, and nothing is added to that.

Changes

If this policy changes in a way that affects what leaves your device, the app will say so rather than relying on you re-reading this page.

Who is responsible, and how to reach us

Data controller: Dataiza s. r. o., Slovakia
Contact: development@dataiza.sk