EUKOSMOS Connector — Privacy Policy

The browser extension that saves papers into your local EUKOSMOS library. Applies to version 1.3.x. Last updated: 22 July 2026. This policy is written directly from the extension's manifest and source code — every permission and every data flow is listed below.

The short version: the extension has no servers. Paper detection is local parsing inside your browser; nothing is captured for saving until you click Save, and the only automatic network requests carry the paper's id to the EUKOSMOS application on your own computer (localhost). We collect nothing — no analytics, no telemetry, no accounts, no cookies, no browsing history.

1. What the extension does

One thing: it detects the academic paper on the page you're viewing and, when you click a button, saves that paper (metadata and, where openly accessible, its PDF) into the EUKOSMOS app running locally on your machine.

2. What data the extension touches, and where it goes

Read from the current page (automatically, in your browser)

On each page load the extension reads standard bibliographic markup: citation_* (Highwire Press) and dc.* (Dublin Core) meta tags — title, authors, journal, year, abstract, DOI, ISSN, volume/issue/pages, language, PDF link — plus DOI or arXiv patterns in the page URL and DOI/PDF links present on the page. This parsing happens entirely inside your browser. Pages with no paper markup produce nothing.

Sent to YOUR local EUKOSMOS app (127.0.0.1 / localhost) — never anywhere else

Automatically, when a paper is detected: the paper's DOI and/or arXiv id is sent to your local app (/api/ext/check-doi) so the toolbar badge can show whether it's already in your library. The extension also probes local ports (/api/ext/ping) to find your running app. These requests carry the paper identifier only, and they go to your own machine.
Only when you click a save button: the full metadata listed above, the page URL it came from, and — where openly accessible — the paper's PDF are sent to your local app, which stores them in your local library folder.

Fetched from the web (only when you click)

When you save a paper with a PDF, the extension downloads that PDF from the publisher or repository hosting it (e.g. arxiv.org) — the same request your browser would make if you clicked the PDF link yourself — and hands it to your local app.

Stored in your browser

Preferences (extension sync storage): target section, My Papers status, theme, optional port override. If you have Chrome's own profile sync enabled, Chrome replicates these few preference values across your devices through your Google account — that is Chrome's mechanism under your control, not ours; no paper data is stored there.
Per-tab paper state (session storage): the metadata detected on each open tab, cleared automatically when the browser closes.

One local courtesy marker

On EUKOSMOS's own local pages (localhost / 127.0.0.1), the extension sets a page attribute announcing its version so the app can show that the Connector is installed. This stays in the page; nothing is transmitted.

3. What we collect

Nothing. The extension has no backend. No data of any kind reaches us or any third party: no analytics, no telemetry, no crash reporting, no browsing history, no personal information, no accounts, no cookies. The only network destinations are your own machine and — on your explicit click — the publisher hosting the PDF you asked to save.

4. Every permission, and why it's needed

Permission	Why the extension needs it
storage	Remembers your preferences (target section, My Papers status, theme, optional port) and short-lived per-tab paper state.
activeTab	Lets the popup identify the tab you're looking at so it can show and act on the paper detected there.
http://localhost/* http://127.0.0.1/*	The extension's only data destination: the EUKOSMOS app running on your own computer. Used to find the app and to deliver papers you save.
Access to all websites (<all_urls>)	Two reasons. (1) Detection: papers live on thousands of publisher, journal and repository domains — an unbounded set, so a fixed domain list is impossible (the same model reference-manager extensions use). The content script only reads bibliographic meta tags and DOI/PDF links. (2) PDF download: when you save a paper, its PDF is hosted on whatever domain the publisher uses.

The extension deliberately does not request: tabs (no access to your other tabs' URLs or titles), history, cookies, downloads, or scripting. It loads no remote code.

5. Changes and contact

If a future version changes any of the above, this page and the version note will be updated before that version ships. Questions: [email protected].