You remember the idea but not the words — so the search box comes up empty. **Semantic search is the technology that finds a note by what it means, not by the exact words it contains.** It's why an app can return the right note even when your search shares zero words with what you wrote.
Here's the everyday version: last month you jotted "why I keep ordering takeout at 9pm." Today you search decision fatigue — same idea, completely different words — and a good notes app still hands you that note. Let's build up to how, starting from the kind of search you already know.
First, the kind of search you already know: keyword search
Open most apps and search for "dog." The app scans your text for the literal letters d-o-g and returns every note that has them. That's keyword search (also called full-text search): match the characters, return the hits. It's fast and it's been the default for decades.
Its weakness is also its rule: it only finds the exact word. Search "dog" and you'll miss the note that says "puppy," "my golden retriever," or "the new pup." Same idea, different letters — invisible to keyword search.
The new idea: turning meaning into numbers
Semantic search solves this by first translating text into meaning, and it does that with something called an embedding.
An embedding is just a list of numbers that represents the meaning of a piece of text. Think of it as a set of coordinates. Just as "40.7° N, 74.0° W" is a point that locates New York City on a map, an embedding is a point that locates a sentence on a giant "map of meaning." A computer model reads your text and places it at a point on that map.
The magic is in where it places things: text with similar meaning lands close together. "My golden retriever" sits right next to "the new pup," even though they share no words. "Decision fatigue" sits next to "why I keep ordering takeout at 9pm," because the model learned they're about the same thing.
How a search actually runs
Now the search itself is simple — it's a distance check on that map:
- You type a query: "decision fatigue."
- The app turns your query into an embedding — a point on the map.
- It measures which of your notes sit closest to that point.
- It returns the nearest notes, ranked by closeness.
No exact word ever had to match. The app found your takeout note because its meaning lives next door to your query's meaning. That's the whole trick: search by proximity of meaning instead of matching of letters.
A worked example
Say your notebook has three notes:
- A: "Why I keep ordering takeout at 9pm — too drained to cook."
- B: "Best dog parks near the office."
- C: "Standing desk made my back feel better."
You search: "willpower runs out at the end of the day."
Keyword search finds nothing — none of those exact words appear in any note. Semantic search turns your query into a point on the map, checks distances, and ranks note A first: "drained," "9pm," and "willpower running out" all cluster in the same neighborhood of meaning. B and C are far away, so they're skipped. You get the note you actually wanted, phrased nothing like your search.
Why this matters for note-taking
The whole promise of writing things down is being able to get them back. But you rarely remember a note in the same words you wrote it — you remember the gist. Keyword search punishes you for that. Semantic search forgives it. It turns a pile of notes you can't re-find into a searchable second brain, and it's the same engine behind "show me related notes" and chatting with your own library — because "related" and "relevant to my question" are both just closeness on the map of meaning.
Try this
Pick something you know you wrote about but can't quite word — a worry, a recurring idea, a half- formed plan. Search for it in your own words today, not the words you think you used back then. With keyword search you'll often come up empty; with semantic search the right note should surface. In JustJot.ai, your notes are embedded automatically, so searching by meaning — and asking your library questions in plain language — works out of the box. The thing you half-remember is closer than you think.