Why can't Chrome search inside my bookmarked pages?

Because Chrome never stored them. A bookmark records the address, the title the tab had when you saved it, the folder and the date — nothing else. Searching those three fields is all the data there is to search. Fetching and indexing every saved page would be a much larger job, and one Chrome has never taken on.

Does searching bookmarks by content need an extension?

Inside the browser, yes — no browser can do it on its own, so anything that reads the pages is an extension with permission to do so. Outside the browser it does not: a desktop app can read the bookmarks file the browser already keeps on your Mac and work from there, without an extension and without touching the browser.

What is the difference between full-text search and search by meaning?

Full-text search keeps every word of every page and matches your words against them, so it finds an exact phrase and misses a paraphrase. Search by meaning keeps a description of each page and compares meanings, so it finds a paraphrase and misses an exact phrase. They fail in opposite directions, which is why the right one depends on how you search.

Guide

Searching bookmarks by what is inside them

No browser searches the contents of a bookmarked page. Chrome, Brave, Edge, Arc and Vivaldi all index three things — the title the page had when you saved it, its address, and the folder you put it in — and match your words against those. If you remember a sentence from the article but not its headline, none of them will find it. What follows is why that is, the four ways round it, and what each one costs.

Updated 28 August 2026

What your browser actually searches

A bookmark is a very small record. When you press ⌘D the browser writes down the address, the text that was in the tab title at that moment, the folder, and the date. That is the whole of it — the page itself is never stored, and nothing about it is read again unless you open it.

So bookmark search has three fields to work with, and it matches your words against them literally. Search "drone" and you get the bookmarks with "drone" in the title or the URL. A page called "Pavo20 mount, v3 remix" is about a drone part and contains the word nowhere.

The title is a snapshot. If the site later renamed the article, your bookmark still carries the old name — and if the tab said "Loading…" when you saved it, that is what got written down.
The URL is often meaningless. Half the web addresses you have saved are a domain and a string of digits.
The folder is a decision you made in a hurry, about a search you would make years later.

None of this is a bug. It is what makes bookmarking instant and independent of the network. The cost lands later, and only for people who save a lot.

Why a folder tree stops working at a few hundred links

Folders answer one question well: "where did I file it". That question is only useful if you can remember the answer. At thirty bookmarks you can. At six hundred you are guessing, and the guess is between four folders that could all plausibly hold the thing.

The deeper problem is that filing happens at the wrong moment. You save a link in two seconds while reading something else, and the folder you pick reflects why the page mattered that afternoon. Two years later you want it for a different reason, and you are searching for the reason you have now, not the one you had then.

This is why the usual advice — prune them, nest them, tag them properly, declare bookmark bankruptcy — never sticks. All of it is work done at save time to serve a search you cannot predict.

The four ways round it, and what each one costs

There are four honest approaches, and which is right depends entirely on what you are trying to find.

A full-text extension. It fetches every bookmarked page and indexes every word, so an exact phrase you half-remember will find it. That is the strongest option if you search for phrases. It also means a browser extension with permission to read every page you saved, and an index that grows with the length of the pages rather than the number of them.
A web archiver. Tools of the ArchiveBox and Linkwarden kind keep a copy of the whole page, which survives the site going down. That is a different and larger job than search, usually with a server to run.
A read-later app. Pocket-shaped services store the article text and search it. They are built for a queue you intend to read, not a library you accumulated, and most of them want the collection in their cloud.
Search by meaning. Instead of storing the words, store what the page was about, and match a question against that. This finds a page whose title shares no word with your question — and it will not find an exact phrase, because the phrase was never kept.

The first three all rest on the same idea: keep more text, then match words against more text. The fourth changes what is being matched.

What searching by meaning adds, and where it stops

Searching by meaning reads each saved page once, writes a line saying what it is, and turns that line into a list of numbers that can be compared with the numbers made from your question. Two pieces of text about the same subject land near each other even when they share no words at all. That is how "that 3D-printed part for the drone" finds a bookmark titled "Pavo20 camera mount and antenna holder".

It also works across languages, because the comparison is on meaning rather than spelling: a question typed in Portuguese can find a page written in English.

Now the limit, which matters more than the promise. What is kept is a description of the page, not the page. If you are hunting for a specific sentence, a code snippet, a price, or a name that appeared once halfway down an article, searching by meaning will not find it and a full-text index will. Choose by which of those two you actually do.

Questions

Why can't Chrome search inside my bookmarked pages?: Because Chrome never stored them. A bookmark records the address, the title the tab had when you saved it, the folder and the date — nothing else. Searching those three fields is all the data there is to search. Fetching and indexing every saved page would be a much larger job, and one Chrome has never taken on.
Does searching bookmarks by content need an extension?: Inside the browser, yes — no browser can do it on its own, so anything that reads the pages is an extension with permission to do so. Outside the browser it does not: a desktop app can read the bookmarks file the browser already keeps on your Mac and work from there, without an extension and without touching the browser.
What is the difference between full-text search and search by meaning?: Full-text search keeps every word of every page and matches your words against them, so it finds an exact phrase and misses a paraphrase. Search by meaning keeps a description of each page and compares meanings, so it finds a paraphrase and misses an exact phrase. They fail in opposite directions, which is why the right one depends on how you search.