Internal project · Productivity tool Codally internal project

Find any file in a few milliseconds

Windows’ file explorer quickly becomes unusable when you search across large volumes: you type a name and you wait. We built our own explorer for our internal needs, then for those of several partners. It indexes the disk once, then finds files instantly, by name as well as by content.

Up to 8×faster than the Windows explorer
< 50 msto show the first results
By contentsearch inside the files
Millionsof files indexed without slowing down
Explorer interface: search results appearing instantly
At a glance
Project
Codally internal tool
Nature
Windows desktop application
Origin
Internal need, extended to our partners
Core
Disk indexing + search engine
Search
By file name and by content
Stack
C++ (engine) · Python · Windows
The starting point

Finding a file shouldn’t take a minute

On a large tree, Windows’ native search walks the folders on every query. The result: several seconds, sometimes more, for a simple file name — and no realistic way to search inside the content of documents.

For our projects as well as our partners’, we handle thousands of files: code, drawings, documents, exports. We needed a tool that answers at the speed of typing, not the speed of the disk.

  • A native search that scans the disk on every query.
  • Several seconds of waiting on large directories.
  • No realistic search inside file content.
  • An experience that breaks focus every day.

The gain

The difference is measured in milliseconds

At equal volume, on the same machine, search goes from a noticeable wait to an immediate result. The values below are orders of magnitude observed on large directories.

Windows explorerOur explorer
Search by name (large volume)~4 s< 50 ms
Search by contentNot realisticA few seconds
Results displayProgressiveInstant as you type
Index updateContinuous, in the background

Under the hood

Index once, answer instantly

The speed doesn’t come from a trick but from an architecture designed for search: we build an index once, keep it up to date in the background, and every query hits memory rather than the disk.

Disk indexing

The engine reads the system’s file table and builds a compact in-memory index, instead of walking folders on every search.

In-memory search engine

Name queries are resolved against the in-memory index, making results near-instant even across millions of files.

Content search

Beyond the name, the tool opens and analyses file content to find text inside documents.

Incremental updates

The index follows disk changes in the background: created, modified or deleted files stay current without a full re-index.

Architecture diagram: indexing, in-memory index and search engine

Capabilities

What the tool can do

  • Instant search by name

    Results appear as you type, with no waiting time.

  • Content search

    Find a file from a word or phrase it contains.

  • Advanced filters

    Filter by type, size, date or location to target the right file.

  • Large volumes

    Millions of files indexed without degrading responsiveness.

  • Always-current index

    Disk changes are reflected continuously, in the background.

  • Lightweight and native

    A fast Windows desktop app with a C++ core.


The uses

A tool born internally, useful far beyond

Designed first for our own projects, the explorer proved useful everywhere people handle a lot of files.

Technical teams

Find a code file, an asset or an export among thousands of others.

Large document repositories

Navigate trees of drawings, contracts and documents without getting lost.

Partners

Deployed at several partners whose teams were losing time to search.


Tech stack
C++PythonWindows APINTFS indexingMultithreading

Is an internal tool slowing you down?

We build custom tools, engineered for performance where generic solutions stall. Let’s talk about yours.

Discuss a custom tool