How to find a note by describing it

Notes get written fast and named badly. “Untitled”, “Ideas”, “Call w/ J”, or nothing at all. The note you want now might be a half-page about pricing a side project, a list of book recommendations from a friend, or the steps someone talked you through to fix the router. You remember what it was about. The words you used, and which app you wrote it in, are another matter.

This guide covers how people usually look for a lost note, and how to find one in Fabric by describing the idea, starting from a related note, or asking the AI.

How people usually look for a lost note

Most notes apps have a search box that matches words. It works when you remember a distinctive term, like a person’s name or a product. It’s less helpful for a note about “what to do with the spare room” when the note itself says “office? guest bed? sell the desk”.

Many people keep notes in more than one place. There’s a phone notes app for quick things, a work tool like Notion or OneNote, a few documents in a cloud drive, and the occasional email to self. Each has its own search, so a lost note means checking each app in turn.

Others rely on structure: folders by project, a pinned index note, or a method such as PARA or Zettelkasten. These systems hold up well for people who maintain them. When a busy month interrupts the habit, the notes from that month are the ones that end up hard to find.

Step 1: Get your notes into one place

Write new notes in Fabric’s editor, which handles markdown. For text written elsewhere, paste it straight into Fabric and it becomes a note. From your phone, use the share sheet in any app to send text to Fabric, and you can forward notes by email to your Fabric email address too.

Notion docs can come in through a connection, which needs a paid plan. Fabric also has import functions for Evernote and Notion if you’d rather move your archive across.

If your notes already live as Markdown or plain text files in a folder, drag the whole folder into Fabric or click Upload a file, which handles large batches at once. With the desktop app installed, you can also move the folder into the Fabric drive like any other folder on your computer.

Recorded thoughts count as notes as well. Start a voice note in any Fabric app and, when you stop, Fabric writes it up. Transcription needs a paid plan.

Step 2: Describe the note in search

Click Search or press Cmd+F (Mac) / Ctrl+F (Windows) and describe the idea: “options for the spare room”, “books Sam recommended”, “how to reset the router”. Search is hybrid, matching by meaning and by keyword, so describing the topic finds notes that never use your search terms.

Every note is processed when it’s saved. Fabric creates embeddings, which give search its sense of meaning, writes a summary, extracts the concepts the note covers, and suggests tags. A three-line note with no title still gets all of that, which is why it can turn up for a description of its subject.

When you do remember a word, quotes give strict matching: “Sam” in quotes returns notes with that exact name. Filter by kind of item to keep results to notes, or by tag if you’ve used any.

Step 3: Start from a note you can find

Ideas tend to come back in several notes over time. If you can find one note on the topic, select it and run a similarity search. Fabric finds items like it across your library, which often brings up the earlier, scrappier note you were after, along with related bookmarks and files.

This suits people who think in connected notes. The Zettelkasten method describes linking notes by hand; similarity search finds related notes you never got around to linking.

Step 4: Ask the AI which note had the idea

For a vague memory, ask the AI assistant: “where did I write down the plan for splitting the garden into zones?” It searches Fabric and your connected apps and answers with clickable references to the notes it used. Follow-up questions narrow it down if the first answer is close but not right.

If you remember which project a note belonged to, open that folder and open the assistant there. It limits itself to the folder and analyzes everything in it. You can also right-click a folder and add it to the assistant’s context.

When the assistant pulls together something useful, such as every note where you mentioned a supplier, save the answer as a note. That gives you a summary to find next time instead of a scattered set of fragments.

Step 5: Let collections gather related notes

If a topic keeps coming up, set up a smart collection for it. Smart collections fill themselves based on content, so new notes on that subject collect there without filing. Folders, tags, kanban and canvas are all available and all optional.

For more on writing notes that are easy to come back to, see note-taking basics.

Frequently asked questions

How do I find a note when I don’t remember any words from it?

Describe what the note was about in Fabric search. It matches by meaning, using embeddings and a summary Fabric creates for each note.

Can I find notes that don’t have a title?

Yes. Fabric processes the content of every note, so an untitled note can be found by its subject.

Can I search notes from other apps?

Paste notes into Fabric, send them from your phone’s share sheet, or forward them by email. Notion can be connected on a paid plan, and the knowledge search panel then searches it with everything else.

Can Fabric import my Evernote or Notion notes?

Yes. Fabric has import functions for Evernote and Notion, as well as Pocket, Omnivore and browser bookmarks files.

How do I find notes related to one I already have?

Select the note and run a similarity search. Fabric returns notes, files and bookmarks with similar meaning.

Will the AI tell me which note its answer came from?

Yes. Answers include clickable references to the notes and files the assistant used.

Do I need to tag or file my notes?

No. Search works without any organizing. Tags, folders and smart collections are there if you want them.

Can I search my notes on my phone?

Yes. Search works in the iOS and Android apps, with the knowledge search panel in the bottom drawer. Download them from the mobile page.