How SeeNote works
Last updated: August 31, 2026
SeeNote asks you to install a browser extension and read securities filings through it. That deserves more than a promise, so this page shows the actual permissions, the actual click path, and how to check both yourself.
The permission model
These are the permission declarations from the extension's manifest.json, reproduced in full; only the line wrapping differs:
"content_scripts": [
{
"matches": [
"https://www.sec.gov/Archives/edgar/data/*",
"https://www.sec.gov/ix*",
"https://www.sec.gov/ixviewer*"
],
"all_frames": true,
"js": ["content-script.js"],
"run_at": "document_idle"
},
{
"matches": ["https://maya.tase.co.il/*"],
"js": ["maya-content.js"],
"run_at": "document_idle"
},
{
"matches": [
"https://mayafiles.tase.co.il/ixbrl/*",
"https://mayafiles.tase.co.il/translated_ixbrl/*"
],
"js": ["ixbrl-content.js"],
"run_at": "document_idle"
},
{
"matches": ["https://data.fca.org.uk/*"],
"js": ["nsm-content.js"],
"run_at": "document_idle"
}
],
"permissions": [
"storage",
"alarms",
"contextMenus",
"activeTab"
],
"host_permissions": [
"http://*/*",
"https://*/*"
]
What each line means:
content_scripts.matches. The filing-reading code injects into four filing surfaces and nothing else: SEC filing pages and the inline XBRL viewer on sec.gov; the Tel Aviv Stock Exchange's Maya site; the two paths onmayafiles.tase.co.ilwhere Maya publishes filings rendered as iXBRL; and the FCA data site that serves the National Storage Mechanism. It runs on no other website: not onwww.tase.co.il, not onwww.fca.org.uk, not on the rest ofmayafiles.tase.co.il.storage. Keeps your settings and local resolution counters on your own device.alarms. A once-a-day timer for the optional usage-count upload; unless you turn that setting on, the timer fires and does nothing.contextMenus. Adds the "Open in SeeNote Reader" entry to the right-click menu for PDF links.activeTab. Lets the toolbar button read the current tab's URL, so a PDF you already have open can move into the Reader.host_permissions. The broad-looking one, so it gets the honest explanation: it exists solely so the Reader can fetch the PDF(s) of the one report you explicitly ask it to open, from whatever site hosts it. Usually that is one file. On Maya it is the report’s attachments — a Tel Aviv periodic report is one submission split across several PDFs, and a reference in the directors’ chapter routinely points into the financial statements, so the Reader has to hold all of them; the fetch is capped at 20 and every one comes from the filing’s own host. It is never used to read pages in the background; the filing-reading code still runs only on the four patterns above.
Where SeeNote works
SeeNote reads three markets, and each one has its own way in.
- SEC EDGAR. Open a 10-K or 10-Q on sec.gov, in the classic HTML view or in the inline XBRL viewer. References become clickable on the page itself; there is nothing to press first.
- Tel Aviv Stock Exchange. On a Maya report page an "Open in SeeNote" button appears beside each filing document, and opens that report in the Reader — the clicked attachment together with the report’s others, because a Tel Aviv periodic report is one submission split across several PDFs and a reference in one chapter routinely points into another. Filings Maya has already rendered as iXBRL are read in place, the same way EDGAR pages are.
- United Kingdom. On the FCA's National Storage Mechanism the same button appears beside each document in the search results, and one button opens exactly one document — deliberately, and unlike Maya: an NSM results page is a search result whose consecutive rows are routinely different companies and different years under the identical heading "Annual Financial Report", so assembling a session from it could let a reference resolve into the wrong filer. A UK annual report is a designed document rather than a filled-in filing template, so its note headings vary far more than EDGAR's; where SeeNote cannot work out which note a reference means, it says so rather than guess.
- Any filing PDF. The Reader opens a PDF you point it at — the toolbar button, a right-click on a PDF link, or drag-and-drop — whatever site it came from. Text-layer PDFs only: a scanned one is detected and reported, never guessed at.
SeeNote does not read Companies House statutory accounts. Companies House re-renders those filings as page images with no text layer, so there is nothing on the page for SeeNote to find, and it does not pretend otherwise.
What happens when you click a reference
You open a 10-K on sec.gov. The page is already fully downloaded by your browser; SeeNote waits until the document is idle, then parses it in place: it finds the notes to the financial statements, works out where each note begins and ends, and wraps every cross-reference like "see Note 12" in a clickable element. A Maya filing or a UK annual report open in the Reader takes the same path, run over a PDF's own text layer instead of a web page's DOM.
When you click one, SeeNote resolves the reference against the index it built and clones the note's content, the exact nodes already sitting in the page, into a popup next to the reference. No request is made. The popup is not a fetched copy, an excerpt, or a rendering of our version of the note; it is the filing's own DOM, duplicated in place. In the Reader, where the document is a PDF, the popup is the printed page's own pixels instead: the exact region of the page, cropped, never re-typeset. Close it and you are exactly where you left off.
Because everything operates on the page you already loaded, popups open instantly and keep working with your network disconnected.
You don't need a reference to reach a note. Press S (or click the on-page button) to open the notes index: a keyboard-navigable list of every note in the filing, in document order. Choose one and it opens in the same popup, cloned from the page in place, still without a request and still verbatim.
Marking a filing yourself
Reading is not only following references, so SeeNote also lets you keep what you found. Select a passage on an EDGAR page or a Tel Aviv iXBRL filing, or drag across a line in the Reader, and pick one of four colours; attach a note of your own if you want one. The mark is anchored to the filing's own text rather than to a pixel position, so it comes back where you left it the next time you open the document. Every mark collects in the sidebar's Highlights tab — filter them, jump back to the page, or copy them all out as text, each quote with your note beneath it. Marks are free on every filing, before and after the trial, and what they store — the quote, your note, the filing's title and address, and when you last changed it — is written to Chrome's local storage on that device and goes nowhere else.
What never leaves your machine
The privacy policy states this as policy; here it is as architecture:
- No servers in the reading path. SeeNote never transmits filing text or page content, to us or to anyone else, under any setting. It does not send which filings you opened either. There is no account and nothing to log into.
- Statistics stay local. The extension counts how many references it resolved or missed, and stores those counters in Chrome's local storage on your device.
- Telemetry is off by default. One optional setting, disabled until you enable it, sends aggregate counters once a day, for example "42 references resolved". No URLs, no tickers, no filing text, no personal data. Leave it off and the extension makes no network requests at all while you read.
- PDFs are read, not uploaded. A PDF you open in the Reader is fetched by your browser and parsed locally. Nothing about which PDF you opened is sent anywhere.
- Your marks stay put. Highlights and the notes you write on them are stored unencrypted in Chrome's local storage on the device you made them on. They are never synced to another computer, never transmitted, and they are removed with the extension.
Why there's no name on this site
SeeNote publishes no personal data about anyone, including whoever built it. That is the extension's privacy principle applied in both directions, not a hedge. The permissions above and the code they gate are the credentials; they are checkable, and a name is not.
Verify it yourself
- Check the permissions. After installing, open
chrome://extensions, find SeeNote, and click Details. The permission list Chrome shows you is the one above; Chrome enforces it regardless of what any website says. - Read the code. Enable Developer mode on
chrome://extensionsand every installed extension's files are plain JavaScript on your disk, open to inspection. - Watch the network. Open DevTools, switch to the Network tab, and read a filing with SeeNote active. Filter by the extension's requests: with usage counts off, there are none.
If you find behavior that contradicts anything on this page, tell us at hello@seenote.co. That report outranks every other inbox we have.