Bulk File Rename Tool

Bulk file rename is the operation of changing a group of file names in one pass according to a set of rules. Typical jobs: turning camera dumps like IMG_1234.JPG into 2026-06-trip-001.jpg, prepending @2x_ to design exports, sequencing scanned receipts by invoice date. macOS Finder's built-in rename covers Find & Replace and a name + counter format but cannot rewrite extensions; Windows Explorer's Tab-rename only adds simple sequence numbers (PowerToys PowerRename adds regex but needs a separate install). Anything more involved usually goes to desktop apps like ReNamer or Bulk Rename Utility, or to hand-rolled PowerShell / shell scripts. This tool exposes each operation — prefix/suffix, numbering, find & replace, case, extension — as a visual control with live preview and duplicate detection, so you can verify the result before downloading. The whole flow runs in browser memory, so filenames containing project codes, client names, or personal information stay on your machine.

The whole tool runs on the browser's File API, Blob, and URL.createObjectURL — every byte stays in memory. When you drop or pick files, the browser hands JavaScript File objects (Blob subclasses) to the JS heap; the tool only reads each file's `name` property for renaming, never touching the binary content. Rules apply in a fixed order to avoid interference: first the base name is determined (user-supplied or original base), with the extension split at lastIndexOf('.') (dotfiles like .gitignore are treated as base only, no extension); then find/replace runs as String.prototype.split + Array.prototype.join (no regex — avoids unintentional special-character matching); next the sequence number is computed via `String(start + i).padStart(pad, '0')` and placed as prefix/suffix/replacement of the base; the prefix and suffix are then concatenated around the result; case transformation applies to the whole prefix+base+suffix string at once; the extension is processed independently (keep/upper/lower/replace). Duplicate detection counts each new name in a Map<string, number> during preview; names appearing more than once enter a duplicateSet that the UI highlights in red. Clicking "Download as ZIP" while the set is non-empty surfaces an error toast and aborts the pack — the original list is preserved so you can adjust the rules and try again. This runs inside useMemo([files, opts]), so rule edits incrementally recompute without re-rendering the file list. Single-file download wraps the File in URL.createObjectURL, triggers a synthetic `<a download>` click, and revokes the URL immediately. ZIP packing dynamically imports JSZip on first click (~80 KB, no first-paint cost), reads each file's arrayBuffer(), writes it under the new name, and calls generateAsync({type:'blob', compression:'DEFLATE', compressionOptions:{level:6}}). DEFLATE is the ZIP default; level 6 balances ratio against speed. Rule computation is O(n) in file count; pack time scales with total bytes. Browsers typically read File contents lazily — arrayBuffer() is when bytes actually leave disk — so even thousands of queued files cost almost nothing during preview, with the memory peak only on "Download as ZIP".

• lastIndexOf('.') splits the extension; if the only dot is at index 0 (e.g. .gitignore) the file has no extension and the leading dot stays in the base name
• Find & replace uses String.prototype.split + Array.prototype.join instead of replaceAll, avoiding accidental interpretation of replacement-string tokens like $1 or $& in the replacement field
• padStart(pad, '0') generates the zero-padded number; pad=0 outputs the raw number for cases where alignment isn't needed
• Case transformation runs after prefix/suffix are joined, so it normalizes the whole base segment at once; the extension is handled independently
• Duplicate detection runs inside useMemo with a Map; O(n) recompute, and the UI only swaps a className for highlighted rows — no DOM rebuild
• Blob URLs are revoked immediately via URL.revokeObjectURL to free the URL allocation pool, since reserved object URLs accumulate quickly under heavy use
• JSZip is loaded via dynamic import only when packing is triggered, so the ~80 KB code never affects first-paint or LCP
• DEFLATE level 6 is the browser-side sweet spot; level 9 adds 1–2% compression at 2–3× the time and barely helps for already-compressed images/video