Large folders
The app is built to stay responsive on folders with tens of thousands of files.
Rendering
The sidebar tree only renders the rows currently on screen, so scrolling through a folder with thousands of entries is as fast as scrolling through a small one.
Scan limits
Scanning a folder stops after 50,000 subfolders. Past that point the sidebar marks the folder as incomplete, with a spinner and a live count of how many directories it has scanned so far.
While a folder is in this incomplete state, moves and deletes that happen inside the part of the tree it hasn't reached yet go unnoticed until a complete scan runs.
Content indexing
Content search indexes files as they are opened, in the background. A file larger than about 2 MB is not content-indexed — it can still be found by name, but a search for text inside it will not match. See Search and frontmatter filters.
Gitignored trees
Show gitignored files has no size cap, so a huge ignored directory (a node_modules, a build
folder) shown this way can slow the sidebar down. Excluding such paths in Exclude Settings avoids
this. It also accepts patterns such as *.json; ** crosses directories only as a whole segment, and rules need a literal character (*.* is allowed). A file row without * or ? in its path and not already excluded offers Exclude from watching, which adds
that one file to the rules. Covered tabs close, saving the active file first; a failed save keeps
its unsaved text open. See Folder watch and external changes.
Reconciliation on a large folder
Reconciliation still runs at launch and on window focus for a large folder, same as any other — see Sidebar and folders. It's the initial deep scan of an unusually large tree, not the ongoing reconciliation, that can leave a folder in the incomplete state described above.
Limits
- 50,000 subfolders per scan; a folder larger than that never fully completes automatically.
- Files over 2 MB are searchable by name only.
- No cap on a shown gitignored tree — exclude it explicitly if it's large.