Guides · measured July 2026
How to search every file on Windows, fast.
Windows ships file search in the Start menu and in every folder window, yet on a real 4.47-million-file drive it takes about 67 seconds to find one file. Mac users are baffled by this, because Spotlight answers as fast as they can type. The difference isn't speed. It's coverage. Here is every way to search files on Windows, why the built-in one is slow, and how to make it answer in the microsecond range.
TL;DR · the short version
Key takeaways
- Windows Search indexes only a slice of your drive: your profile folders, the Start menu, mail. Ask for a file anywhere else and it walks the disk: about 67 seconds on a 4.47-million-file machine.
- A whole-drive index that stays always warm answers the same filename query in 85 µs on a single file (139 µs median), roughly 480,000× faster than Windows Search and 412,000× faster than VS Code's picker.
- Whole-drive content search lands in the 1 to 10 ms range: the hunt that takes ripgrep 93.8 s drops to 16 ms (58,625×).
- For an AI agent the gap is widest: Claude Code's 6 m 57 s, 71-call file hunt becomes 16 ms and one call, and the ~58% of session tokens spent searching drop toward zero.
- It stays out of the way: ~44 MB at rest, searchable ~1 ms after you save. Everything local is free, forever: the app, whole-machine search, MCP. Hosting starts at $5.99/mo.
Your options, honestly
Four ways to find a file, and what each one costs.
Every Windows machine gives you the first three for free. They share one trait: none of them keeps a complete, current index of your whole drive, so on a large disk they either miss the file or walk the filesystem looking for it. Only the last keeps the whole drive pre-indexed.
| The method | Where it is | What it does | Time · 4.47M files |
|---|---|---|---|
| Start menu search | The taskbar search box | Checks a partial index, merges web results, walks on a miss | 67 s median |
| File Explorer search | The box in any folder window | Uses the index if that folder is indexed; otherwise walks the tree live | seconds → minutes |
| The command line | PowerShell · dir /s · findstr | Walks the whole tree on every run: no index at all | grows with the drive |
| Interlinked | Taskbar app + agent tools | Whole drive pre-indexed and kept current: direct lookup | 139 µs median |
Methodology: Ryzen 9 9950X3D · 64 GB DDR5 · NVMe · Windows 11 · 4,470,000 files. The Windows Search figure is a median for a file that lives outside the default indexed folders; the Interlinked figure is the median filename lookup across 20 queries on the same corpus (85 µs on a single file). The middle two vary with folder size, so they are given as ranges rather than a single number.
What the built-in actually indexes
Windows Search covers a slice. Your files live in the rest.
By default, Windows Search indexes a subset of your drive: your profile folders (Documents, Desktop, Downloads), the Start menu, mail. Everything else (Program Files, your code repositories, other drives, the vast middle of the disk) is not in the index. Ask for a file that lives outside the slice and Windows falls back to walking the filesystem folder by folder. On 4.47 million files, that walk is where the 67 seconds goes.
You can widen the slice (Windows 11's Enhanced indexing mode will index more of the drive), but the indexer stays throttled so it doesn't bother the rest of the machine, so on a big, busy disk it rarely catches up. The lesson of the picture is simple: coverage, not clever code, is what makes a search instant. An index that covers every file never has to walk.
The walk
A miss doesn't fail. It walks.
A directory walk scales with the size of the disk, not the difficulty of the question. Every tool that lacks a whole-drive index lands on the same wrong side of the same cliff: VS Code's file picker outside its workspace, the command line, an AI agent shelling out to ripgrep. Plotted on a log time axis, the indexed answers and the walked answers don't just differ; they occupy different neighborhoods.
Everything pre-indexed sits in the microsecond-to-millisecond band on the left. Everything that walks the disk (VS Code at 35 seconds, Windows Search at 67, an AI agent that took 6 minutes 57 seconds to hunt down one file) piles up on the right. The empty middle is the whole point: there is no in-between. You are either on the index or you are on the walk.
The anchor · macOS
Spotlight feels instant because it covers everything.
The reason Mac users expect file search to be immediate is that macOS indexes the whole startup volume by default and keeps it current, so a search almost never falls back to a walk. That is the entire trick. Not faster hardware, not smarter ranking: the index simply covers the files you ask for. Windows can be pushed toward the same behaviour, but between a partial default scope, a throttled indexer, and web results merged into the box, it rarely gets there on a large drive.
Windows
A subset by default; walks the rest
seconds → minutes on a big drive
macOS
The whole startup volume, kept current
as fast as you can type
Windows
Every file + contents, always warm
139 µs median
Spotlight proves whole-volume indexing is possible and expected, but it is macOS-only, and it is still built for a human typing into a box. It was never designed for an AI agent making hundreds of file lookups a minute. On Windows, and for agents, coverage plus freshness has to come from somewhere else.
The fast answer
Cover the whole drive. Keep it warm.
Interlinked keeps a live index of every file on the machine (names and contents) and stays current as files change, so nothing ever falls back to a walk. A single filename lookup returns in about 85 microseconds (139 µs median across 20 queries). Content search across the whole drive lands in the 1 to 10 millisecond range. Save a file and its contents are searchable roughly a millisecond later. At rest the whole engine sits around 44 MB, and the index on disk stays under 1% of the drive.
Agents are the worst case (hundreds of lookups per session), so Interlinked hands the same index to any AI client that speaks MCP: one install auto-configures 19 of them (Claude Code, Cursor, Codex, Cline, Continue and more). Same agent, same query: the lookup that took Claude Code 6 minutes 57 seconds and 71 tool calls comes back in 16 milliseconds and one call, and the session tokens that used to go to hunting for files drop from roughly 58% toward zero.
Do this today
Make Windows file search instant.
Install once. One signed installer sets up the whole-drive index and auto-configures 19 AI clients: Claude Code, Cursor, Copilot, Codex, Windsurf, Zed, Cline and the rest. No folder-by-folder indexing settings, no JSON editing.
Let it cover everything. Filename search works immediately; content search fills in behind it. Because the index covers the whole drive, no query ever falls back to a walk: the thing that made the built-in slow simply can't happen.
Scope when you know where to look. Whole-machine when you don't know where a file lives; scoped to a folder when you do. Both return in the microsecond-to-millisecond range. Scoping just makes it more precise.
Search from the taskbar, or from your agent. Type in the app for yourself, or let Claude Code, Cursor, and Codex query the same index over MCP. One warm index, every searcher on the machine.
Everything local is free forever.
The full product (the same index exposed to every AI client over MCP): free forever, no card. Hosting your projects starts at $5.99/mo. Download for Windows.
Common questions
Frequently asked questions.
Why is Windows Search so slow at finding files?
By default it indexes only a slice of your drive: your profile folders, the Start menu, mail. Ask for a file outside that slice and Windows falls back to walking the filesystem folder by folder, which on a 4.47-million-file drive takes about 67 seconds. It's a coverage problem, not a speed one.
How much faster is a whole-drive index, really?
A single filename lookup returns in about 85 µs (139 µs median across 20 queries), roughly 480,000× faster than Windows Search and 412,000× faster than VS Code's picker. Whole-drive content search lands in 1 to 10 ms; the hunt that takes ripgrep 93.8 s drops to 16 ms (58,625×).
Does it stay fast after files change?
Yes. The index is kept current as files change, so save a file and its contents are searchable roughly 1 ms later: nothing ever falls back to a walk. At rest the whole engine sits around 44 MB.
Can my AI agent use the same index?
Yes, one install auto-configures 19 MCP clients (Claude Code, Cursor, Codex, Cline and more). The file hunt that took Claude Code 6 m 57 s and 71 tool calls comes back in 16 ms and one call, and the ~58% of session tokens spent hunting for files drop toward zero.
What does it cost?
Everything (the app and the same index exposed to any AI client over MCP) is free forever, no card. Hosting starts at $5.99/mo.