What is semantic search?
You saved something months ago and now you only remember what it was about — not its title. Semantic search is built for exactly that gap. Here is how it works, without the jargon.
Search by meaning, not words
Keyword search matches the letters you type against the letters on the page. Semantic search matches the idea. Describe "the article about staying focused" and it can surface a piece on deep work that never used your exact phrase.
How it works, briefly
Text is turned into a list of numbers — an embedding — that captures its meaning. Things that mean similar things end up with similar numbers. When you search, your query is turned into numbers too, and the closest matches come back. No exact wording required.
Why it changes bookmarking
A small list is easy to scan. A library of hundreds is not — unless you can ask for things the way you actually remember them. In Notabe, every save is embedded so semantic search finds it by meaning, which is what keeps a growing library from becoming a graveyard.
Frequently asked
What is semantic search in simple terms?
Search that understands meaning. Instead of matching the exact words you type against the exact words on a page, it matches the idea — so "that piece about focus and deep work" can find an article even if it never used those words.
How is it different from keyword search?
Keyword search looks for literal word matches. Semantic search compares meaning using numerical representations of text called embeddings, so related concepts score as close even when the wording differs.
Why does it matter for bookmarks?
Because you rarely remember the title of something you saved months ago — you remember what it was about. Semantic search lets you describe it and still find it, which is what makes a large library usable.