LinkedOut

LinkedOut — hide Promoted, Suggested, AI slop & "People you may know" from your LinkedIn feed. (Formerly "LinkedIn Feed Filter".)

A free Chrome extension that hides Promoted (ads) and Suggested posts from your LinkedIn home feed, so you only see content from the people you actually chose to follow.

It hides posts that are already on the page (by setting them to "invisible"). It is not an ad blocker — it doesn't touch your network traffic, cookies, or anything outside the LinkedIn feed.

Is this safe? Can I trust it? (read this first)

Short answer: yes, and you don't have to take our word for it — you can see everything it does.

• It can't see anything except LinkedIn. Chrome only lets it run on linkedin.com. It is completely blind to your email, your banking, every other tab, and the rest of the web.
• It never sends data anywhere. There is no "phone home." The code contains no fetch, no network calls, no analytics, no telemetry, no tracking — by design. Your data never leaves your computer because the extension never transmits anything, full stop.
• It collects nothing about you. The only thing it remembers is your two on/off switches. No account, no sign-up, no profile, no history.
• The only permission it asks for is "storage" — that's just so it can remember whether your switches are on or off. It does not ask to read your tabs, your cookies, your passwords, or your browsing history. (Chrome shows you the exact permissions before you install.)
• It's small and readable. The whole thing is about 740 lines of plain JavaScript with no hidden packages, no "build step," and no outside code pulled from the internet — readable end to end in one sitting.
• You can verify it yourself. Run bash tools/audit.sh and it automatically checks that none of the dangerous patterns (hidden network calls, code injection, extra permissions) are present. It currently passes every check.

In plain terms: the worst this extension can do is hide a LinkedIn post you wanted to see. If that ever happens, just flip the switch back off.

---

Install it in 5 minutes (no coding required)

🛍️ Chrome Web Store listing: coming soon. Once it's live you'll get a one-click "Add to Chrome" button with automatic updates. Until then, install from source using the steps below — it works exactly the same.

You do not need to know how to code, and you don't need any developer tools.

Step 1 — Download the files

1. Go to the top of this GitHub page.
2. Click the green < > Code button.
3. Click Download ZIP.
4. Find the downloaded file (usually in your Downloads folder) and unzip it (double-click it on Mac, or right-click → Extract All on Windows). You'll get a folder named something like LI-BS-Filter-main.
5. Remember where that folder is. Inside it you should see a file called manifest.json — that's the one Chrome needs.

Step 2 — Turn on Chrome's "Developer mode"

1. Open Chrome.
2. In the address bar, type chrome://extensions and press Enter.
3. In the top-right corner, flip on the Developer mode switch.

"Developer mode" sounds scary but just means "let me install an extension from a folder instead of the Chrome Web Store." Nothing else changes.

Step 3 — Load the extension

1. Click the Load unpacked button (top-left).
2. Select the folder you unzipped in Step 1 (the one containing manifest.json).
3. Done — "LinkedOut" now appears in your list, and a small icon appears in your Chrome toolbar.

Step 4 — Use it

1. Go to linkedin.com/feed.
2. Promoted and Suggested posts are hidden automatically.
3. A small control panel appears at the bottom-left of the page — flip any filter on/off, see live counts, or click 👁 Show hidden posts to peek at what was filtered. (You can also click the toolbar icon for the same controls.)

Tip: if you don't see the icon, click the puzzle-piece 🧩 icon in Chrome's toolbar and pin "LinkedOut" so it's always visible.

Keeping it updated

Because you installed it from a folder, it won't auto-update. To get a newer version, download the ZIP again, unzip it over the old folder, then click the 🔄 refresh icon on the extension's card at chrome://extensions.

---

Frequently asked

Will LinkedIn know I'm using this or ban me? The risk is low. LinkedOut only changes what you see — it hides cards that are already on your screen, the same as scrolling past them. It doesn't log in, click, scrape, or talk to LinkedIn's servers, so there's nothing on their end to detect. That said, LinkedIn's User Agreement broadly discourages software that modifies the service, and no extension can guarantee immunity — use it because it improves your own experience, not on a promise from us. (The MIT license is provided without warranty.)

Does it work with uBlock Origin / other extensions? Yes. It only touches the feed's visible posts and doesn't interfere with other extensions.

It hid a post I wanted to see — what do I do? Open the popup and turn off "Hide Promoted" or "Hide Suggested." The post comes right back.

My LinkedIn isn't in English. Already handled — it recognizes the labels in English, German, French, Spanish, Portuguese, Italian, Dutch, Norwegian, Swedish, Japanese, and Simplified Chinese. (Adding more is easy — see Updating the label map below.)

---

Features

• Four independent filters, each its own toggle:
  • Promoted — paid ads (on by default)
  • Suggested — algorithmic posts from outside your network (on by default)
  • Recommendations — "People you may know", "Add to your feed", follow suggestions (off by default — opt in)
  • News & trending — "Suggested News for You" and trending cards (off by default — opt in)
  • Job recommendations — "Jobs recommended for you" and similar job panels (on by default)
• On-page control panel — a small floating widget at the bottom-left of your feed lets you flip any filter and peek at hidden posts ("👁 Show hidden") without opening the toolbar menu
• Preferences persist across sessions and devices (chrome.storage.sync)
• Live badge count of posts filtered this session
• Handles infinite scroll via a debounced MutationObserver
• Resilient detection: matches localized label text, not brittle CSS class names
• Data-driven categories — supporting a new LinkedIn "junk" label is a one-line addition
• Zero data collection — no analytics, no telemetry, no external calls
• No build step, no npm packages, no CDN imports — pure vanilla JS/HTML/CSS

File layout

File	Role
manifest.json	Extension config, permissions, CSP
content.js	DOM observation, label detection, hiding
background.js	Service worker — badge + session counts
popup.html / popup.js / popup.css	Toggle UI and session stats
icons/	Toolbar icons (16/48/128 px)
tools/gen_icons.py	Regenerates the icons (stdlib only, optional)
tools/audit.sh	Automated security/compliance audit

How detection works

For each feed card added to the DOM, the content script inspects <span> elements and compares their trimmed, lowercased text against a localized label map. On a match it walks up to the enclosing feed-card container and sets display:none, tagging it with data-lff-hidden so it can be restored if you toggle the filter off. Detection is intentionally text-based so it survives LinkedIn's frequent DOM/class-name changes.

Privacy & security

• Requests only the storage permission and the linkedin.com host permission — nothing else.
• Strict CSP (script-src 'self'; object-src 'none').
• No eval, no innerHTML writes, no dynamic script creation, no fetch / XMLHttpRequest / WebSocket / sendBeacon.

See PRIVACY.md for the full privacy policy.

Run the automated security/compliance audit (PRD §8.2) any time:

bash tools/audit.sh

Regenerating icons

The icons are committed, but you can rebuild them with the standard library only:

python3 tools/gen_icons.py

Adding labels or whole new filters

Detection is data-driven. Near the top of content.js is a CATEGORIES array; each entry is one independently toggleable filter with an exact list of label strings. To support a new language, add the translated string to the relevant category's exact list. To add a brand-new filter, append a new category object (give it an id, key, label, defaultOn, and exact list) and add a matching toggle in popup.html plus its key in background.js's DEFAULT_PREFS. Matching is case-insensitive and whitespace-trimmed.

The Promoted and Suggested categories are fully localized (English, German, French, Spanish, Portuguese, Italian, Dutch, Norwegian, Swedish, Japanese, Simplified Chinese). The opt-in Recommendations and News categories ship with English labels and are easy to extend the same way.

License

MIT © 2026 Rocky Verma / Kaipability