How to Organize Your Notes and Files Together Using a Graph

Quick answer: In Filarr you connect a note to another note, a file, or a whole folder by typing a wiki-link like [[Project Kickoff]], [[file:contract.pdf]], or [[folder:Clients]] inside the note. Every link you write becomes an edge in the graph view, where notes, files, and folders appear as connected dots you can click, drag, filter, and cluster. The result is a living map of how your work actually fits together, so you find things by following connections instead of remembering which folder you buried them in. It is free for local use, and everything stays encrypted on your machine.

The problem: folders forget, connections don't

Most of us organize digital work the same way we did in 1995: nested folders. Work > Clients > Acme > 2026 > Q1. It feels tidy right up until the moment a file belongs in three places at once. The signed contract is a client thing, a legal thing, and a this-specific-project thing. A folder tree forces you to pick one home and abandon the others. Six months later you are searching for a PDF you know exists, opening five folders, and quietly losing ten minutes of your life.

Notes make it worse. Your meeting notes reference a spec. The spec references a budget spreadsheet. The budget references a decision you wrote down in a daily note. None of those relationships live anywhere except your memory, and memory is exactly the thing you were trying to offload.

A graph fixes this by storing the relationships instead of forcing a single location. Instead of asking "which folder did I put this in," you ask "what is this connected to," and you follow the thread. This guide walks through exactly how to build that in Filarr: how to link notes to notes, notes to files, and notes to folders, how to read the graph once it grows, and the habits that keep it useful instead of turning it into a hairball.

This is a practical how-to. If you want the bigger picture of running your whole life from one connected space, the companion piece on how to organize your whole digital life in one private workspace covers the mindset. Here we stay hands-on.

Step 1: Understand what a link actually is in Filarr

Before you touch the graph, you need to understand the one mechanic everything rests on: the wiki-link. In Filarr a link is a piece of text you type inside a note using double square brackets. There are three kinds, and they map directly onto the three things you want to connect.

[[note title]] links to another note. Example: [[Q1 Roadmap]].
[[file:filename]] links to a stored file. Example: [[file:acme-contract.pdf]].
[[folder:folder name]] links to a whole folder. Example: [[folder:Clients]].

You can also alias a link so the visible text reads naturally while still pointing at the right target: [[Q1 Roadmap|our roadmap for this quarter]] shows "our roadmap for this quarter" but links to the Q1 Roadmap note. And if you want the target rendered inline instead of just referenced, you prefix it with an exclamation mark to embed it: ![[file:diagram.png]].

The matching is by name, and it is case-insensitive, so [[q1 roadmap]] and [[Q1 Roadmap]] resolve to the same note. That is convenient, but it has a consequence worth remembering: two notes with the exact same title are ambiguous. Give your notes distinct titles and this never bites you.

Here is the important part for organization: the moment a link resolves to a real note, file, or folder, Filarr records that connection on both ends. The note knows it points outward, and the target now has a backlink pointing home. That two-way bookkeeping is what makes the graph possible, and it is also what makes the graph honest. Only links you actually wrote show up. Filarr deliberately does not invent connections from folder membership or tags, because artificial edges create fake clusters and drown out the structure you built on purpose.

Step 2: Link your first note to a file

Let's make this concrete with the single most useful link type: a note that points at a file.

Say you had a client meeting. Create a note titled Acme Kickoff Meeting. Write your notes normally. Now, wherever you mention the contract, type [[file:acme-contract.pdf]], using the file's real name as it is stored in Filarr. As you type inside the brackets, Filarr offers suggestions from your existing files, so you rarely have to type the whole name. Pick the contract from the list.

That is the whole action. You now have a note that carries its own attachment context. Open the note in six months and the contract is one click away. Open the file's backlinks and you can see every note that ever referenced it, with a snippet of surrounding text for context, so you remember why it mattered.

This is the pattern that ends the "where did I put that file" problem for good. You stop relying on folder location and start relying on the note where the file was relevant. If you have felt the pain of scattered attachments, the deeper walkthrough on how to stop losing files across apps builds on exactly this habit.

Step 3: Link notes to each other to build a train of thought

Files are the easy win. The real payoff comes from linking notes to notes, because that is where your thinking lives.

Work the way you already think. In your Acme Kickoff Meeting note, when you reference the roadmap, write [[Q1 Roadmap]]. In the roadmap note, when you mention the budget, write [[Acme Budget]]. You are not building an index up front. You are dropping a link every time one idea genuinely leans on another. Over a few weeks these individual links accumulate into a web that mirrors how your projects actually connect.

If the note you want to link to does not exist yet, that is fine. Type the link anyway. When you later create a note with that title, the link resolves automatically and the connection appears. This lets you capture at full speed and fill in the details later, which is the core move behind building a second brain that keeps notes, files, and a linking graph in one place.

Two habits make note-to-note linking pay off:

Link on mention, not on schedule. The instant you type the name of another topic that has (or deserves) its own note, wrap it in brackets. Don't save linking for a weekly cleanup. The friction of coming back later is exactly what kills the habit.
Prefer many small linked notes over a few giant ones. A note about one decision, one meeting, one concept, linked to its neighbors, is far more useful in a graph than a 3,000-word mega-note that connects to nothing.

Step 4: Open the graph and read it

Once you have a handful of links, open the graph view. This is where the structure you have been building becomes visible.

Every item is a node, and Filarr draws the three types differently so you can tell them apart at a glance:

Notes are circles, marked with an N.
Files are squares, marked with an F.
Folders are hexagons, marked with a D.

The size of a node reflects how connected it is. A note that many things link to grows larger, so your hubs, the central documents everything else hangs off, literally stand out. A node with no connections stays a small dot. This is your first diagnostic: big nodes are load-bearing, isolated dots are orphans that might need linking (or might just be fine on their own).

The graph is fully interactive. Scroll to zoom, drag the background to pan, and use the reset button to fit everything back on screen. Drag any node to reposition it, and it stays pinned where you drop it, so you can arrange your map spatially and it remembers the layout between sessions. Click a note node to open it in the editor. Click a file or folder node and Filarr jumps you straight to it in your files. The graph is not a read-only picture; it is a navigation surface.

Step 5: Use clusters, heat map, and time travel to find meaning

A raw dot-and-line diagram is pretty but limited. Filarr layers three analysis modes on top, and each answers a different question.

Clusters answer what are the natural groupings in my work? Filarr runs community detection (a simplified Louvain algorithm) over your links and colors densely connected nodes together. You did not label these groups; they emerge from how you actually linked things. Turning on clusters often reveals a project, a theme, or an area of your life you did not realize had become a coherent body of work. The legend and the stats bar tell you how many clusters were found.

Heat map answers what have I been working on lately? It recolors nodes by recency, from warm red for notes touched today, through amber for this week, to cool blue for older material. When you open your graph after time away, heat map instantly shows you where the live activity is and which corners have gone cold. It is a surprisingly good "where was I" cue.

Time travel answers how did this grow? A slider lets you scrub back and see your graph as it existed at an earlier date, with notes appearing as they were created over time. It is partly satisfying and partly useful: watching a cluster form is a good way to understand how a project actually developed versus how you remember it developing.

There is also a plain filter box: type a word and the graph highlights matching nodes, which is the fastest way to locate one specific thing in a large map.

Step 6: Create links directly from the graph

You do not have to go back into a note every time you want to connect two things. In the graph itself, hold Shift and click one note, then Shift-click a second note, and Filarr writes the wiki-link into the first note for you and confirms with a small toast. Press Escape to cancel mid-action. If the link already exists, Filarr tells you rather than duplicating it.

This turns the graph into an active thinking tool. You spot two ideas sitting near each other that clearly belong together, you Shift-click, and the connection is real and saved. It is the closest thing to drawing on a whiteboard that also updates your actual notes.

Pitfalls and best practices

A graph rewards good habits and punishes lazy ones. Here is what breaks it and how to keep it healthy.

Duplicate titles create ambiguity. Because links match by name, two notes called Notes or three called Meeting will confuse resolution. Give notes specific titles: Acme Kickoff Meeting, not Meeting.

Match file and folder names exactly. A [[file:report]] link only resolves if a file named report (case aside) exists. If you rename a file, check the notes that referenced it. Filarr can propagate renames through links, but it is worth a glance.

Do not over-link. If every note links to every other note, the graph becomes a hairball and clusters stop meaning anything. Link where there is a real dependency, not out of completionism. A slightly sparse graph is more readable than a dense one.

Beware the orphan pile. Notes with zero links show up as tiny isolated dots. A few orphans are normal (a quick capture, a scratch note). A large field of them means you are capturing but never connecting. Spend two minutes now and then linking orphans into the web.

Very large graphs are capped for performance. Filarr keeps the graph responsive by rendering up to a couple of thousand of your most-connected nodes at once. If you have an enormous vault, use the filter and clusters rather than expecting every single node on screen simultaneously. In practice this ceiling is high enough that most people never hit it.

A graph complements folders, it does not replace them. Keep using folders for coarse storage. Use the graph for relationships and retrieval. The two together are stronger than either alone, which is the same principle behind keeping projects, tasks, notes, and files in one connected workspace instead of scattering them.

Where the encryption fits in

Everything above is about getting organized. It is worth saying, once, where your privacy sits in all this, because it is easy to forget it is even happening.

Every note and every file that becomes a node in your graph lives encrypted on your own disk with AES-256-GCM, each file sealed with its own key. The graph is computed locally, from your local data. There is no server building a map of how your ideas connect, because there is no server involved at all unless you opt into cloud sync, and even then it only ever receives already-encrypted blobs it cannot read. Your knowledge graph, arguably the most revealing picture of how your mind works, never leaves your control. If you want the mechanics, the piece on why per-file encryption matters explains the key hierarchy.

That is the quiet part of the deal: you get the organizational power of a linked graph, and you do not trade away who gets to see it. The encryption is reassurance in the background. The linking is what actually changes your day.

FAQ

Do I have to learn the graph before I can use Filarr? No. Write notes and store files normally. The graph is built passively from the wiki-links you drop in, so it grows on its own. You can ignore the graph view entirely for weeks, then open it and find a map already waiting.

What is the exact syntax for linking? [[Note Title]] for a note, [[file:filename.ext]] for a file, [[folder:Folder Name]] for a folder. Add |alias to change the display text, and prefix with ! to embed the target inline. Matching is case-insensitive and by name.

Can I link to a file, or only to notes? Both, plus folders. That is the point of Filarr's graph: it puts notes, files, and folders on the same map, so a meeting note and the PDF it discusses live one click apart. Files show as squares, folders as hexagons, notes as circles.

What are clusters and did I create them? Clusters are groups of densely interlinked nodes that Filarr detects automatically and colors together. You did not label them; they emerge from your actual links. They are a good way to discover themes in your own work you had not consciously named.

Will linking my notes send data anywhere? No. The graph is computed locally from data that is encrypted on your disk. Cloud sync is optional, and when enabled it only ever stores opaque encrypted blobs, so even then no server can reconstruct your graph. This is the same local-first, zero-knowledge model described in what local-first software actually means.

How is this different from Obsidian's graph? Both are local-first and both use a force-directed graph. The main differences are that Filarr encrypts by default and puts files and folders in the graph alongside notes, not just notes. Obsidian has a far larger plugin ecosystem and a more mature community. The honest side-by-side is in the Obsidian vs Filarr comparison.

Conclusion: start with three links today

You do not build a useful graph by planning it. You build it by linking as you go. So do three things right now: open your most recent note, add one [[file:...]] link to a document it references, one [[another note]] link to a related idea, and one [[folder:...]] link to where the work lives. Then open the graph and watch those three connections appear.

Repeat that reflex every time you write, and within a couple of weeks you will have a map you actually navigate by, one where finding something means following a thread instead of guessing a folder. And the whole thing stays encrypted on your machine, free for local use, yours to keep. That is the calm version of a digital life: organized by meaning, private by default.