ZRename V1.0.0

ZRename

Rename ten thousand files without fear

A bulk file renamer for Linux and Windows. Drop a folder in, stack up rules,
watch a live before/after table with collisions flagged in red, and press Apply.
If it was wrong, press Undo — and every file goes back, including after a
reboot.

zsync.eu/zrename — homepage and downloads

_{Free · open source · no telemetry · no account · works offline}

Preview before you commit A live table with a character-level diff, so you see the exact result of every rule before a single file moves.

Undo that actually works Every batch is journalled to disk before the first rename. Undo a batch you ran yesterday.

Rules that stack Eleven rule types, reorderable and individually toggleable, like an image editor's adjustment layers.

Names from metadata EXIF, ID3, video, PDF and hashes turn a string tool into a librarian.

---

Contents

• Why another bulk rename tool
• How it keeps your files safe
• Screenshots
• Every feature
• Install
• Quick start
• The rules
• Metadata tokens
• Presets
• Filters
• Command line
• Keyboard
• Performance
• Where things live
• Building from source
• Licence

---

Why another bulk rename tool

Renaming a pile of files is a chore that hits everyone — phone photo dumps,
downloaded series, exported invoices, asset folders, scanned documents — and the
existing tools are genuinely bad at it.

Tool	What it gets wrong
Advanced Renamer	Windows only · cluttered interface · closed source
Bulk Rename Utility	Windows only · legendarily overwhelming, fourteen panels at once
PowerToys PowerRename	Windows only · one regex rule · no metadata · no undo journal
krename	KDE-bound · dated · weak preview · no journalled undo
Thunar bulk rename	Three rule types · no stacking · no presets
rename / shell loops	No preview · no undo · one typo from wrecking a folder
ZRename	Linux and Windows · eleven stacking rules · metadata tokens · undo that survives a reboot

Two things stand out on that list.

Almost none of them have a real undo — and when a rule was wrong on 4,000
files, undo is the only feature that matters.

And none of them handle the case-insensitive filesystem trap. Renaming
photo.JPG → photo.jpg on Windows silently fails or clobbers the file unless
the rename goes through a temporary name first. ZRename detects that case and
does the two-phase dance for you, without asking and without mentioning it.

---

How it keeps your files safe

Correctness is the entire product, so the engine is built in that order.

files × rules ──▶  plan  ──▶  validate  ──▶  journal  ──▶  execute
                 (a pure     (collisions,   (written     (two-phase
                  function)   bad names,     and flushed   where a cycle
                              limits)        first)        needs it)

1 · Plan first, touch nothing files × rules → plan is a pure function. Nothing on disk is touched while you type in a rule, so the preview can never half-apply something. 2 · Validate against the real filesystem Collisions — case-folded when the filesystem ignores case — characters the target rejects, Windows device names like CON.txt, name length in the right unit (255 bytes on ext4 versus 255 UTF-16 units on NTFS), and the 260-character MAX_PATH limit.

3 · Journal, then execute The undo journal is written and flushed to disk before the first rename, so a crash mid-batch still leaves a complete record. Swap cycles (a→b, b→a) and case-only changes go through a temporary name; everything else renames directly. 4 · Undo verifies before it moves Each file's size and modification time are checked against the journal first. A file you edited since the rename is reported, never overwritten.

Important

Overwrite is never a default. The conflict policy starts at stop and
report it, and even when you choose overwriting it refuses to settle two
selected files that want the same name — one of them would simply be lost.

Tip

Windows filesystem rules are stored as data, not #[cfg] blocks. An
FsProfile value describes case folding, illegal characters, reserved device
names and length limits, so the NTFS and FAT32 rule sets are exercised by unit
tests on any machine — and re-run against real NTFS on a Windows CI runner.

---

Screenshots

Collisions block Apply Rows that would land on a name another row wants are flagged, sorted to the top, and Apply stays disabled with a plain-English reason. Collision red is used for collisions and for nothing else in the app.	A regex tester that removes the guesswork The editor tests your pattern against the row you have selected and shows the capture groups and the final result as you type. Rules toggle off without being deleted.
Undo, even after a restart After applying, the commit bar becomes a confirmation strip with Undo. Every batch also lands in the history list — and those survive quitting the app.	Light is a first-class theme Long file lists genuinely read better on a light ground, so Bench Light is not an afterthought. Bench, Nord and a high-Contrast theme ship alongside it.
Everything from the keyboard Ctrl K reaches every command, preset, rule and theme without leaving the keyboard.	Settings that explain themselves Four themes, row density, paranoid mode, path-limit handling, and what to do when a file has no value for a token.
Filters that reach the whole folder Extensions, globs both ways, a name regex, a size range and a date range — before any rule runs.	Duplicate-aware Grouped by size first, then hashed, so only real candidates are read. It can untick the copies and leave the first of each group to be renamed.

---

Every feature

Rules Find & replace — plain text or regex with $1 capture groups, case-sensitive or not, first match or all Case — lower, UPPER, Title, Sentence, camelCase, PascalCase, snake_case, kebab-case Insert — at a position, at the start or end, before or after a marker Remove — a character range, a set of characters, a word, every digit, or repeated text Trim — surrounding whitespace, a custom character set, collapse runs of spaces Numbering — start, step, padding, per-folder reset, ascending or descending, sorted by name, natural order, size or date Extension — set, fill when missing, lowercase, uppercase, remove Sanitise — strip illegal characters, collapse spaces, transliterate Unicode → ASCII, trim trailing dots Template — rebuild the name from metadata tokens Move into folders — file results into subfolders derived from tokens CSV mapping — rename from an external old,new list Every rule can be scoped to the name, the extension, or both. Reorder by dragging, by Alt ↑/↓, or from the card menu.

Preview Virtualised table — 100,000 rows scroll without stutter Character-level diff: green for added, struck-through red for removed Per-row status: ok, unchanged, collision, invalid, too long, reserved name, skipped Leading and trailing whitespace shown as a visible · Every filename in a monospace face, so a double space or an l/1 is obvious Symlinks marked — ZRename renames the link, never its target Folders marked, moves marked, case-only changes marked Untick any row to leave that one file alone Filter by text, or show only problems Hide unchanged rows Collisions sorted to the top Draggable, persisted column widths Comfortable or compact row height Drag the table out as CSV, or write it with Dry run

Safety and workflow Journalled undo that survives a restart, with a history list per batch Undo verifies size and modification time before moving anything Two-phase renames for swap cycles and case-only changes Conflict policy: stop · skip · suffix (2) · overwrite Warns before Apply when the target filesystem has rules your names may not survive Paranoid mode hardlinks every original into a temp directory first Presets as plain, shareable TOML, with import and export Watch a folder and apply the stack to whatever arrives Duplicate detection that can untick the copies Files with a missing token can be left alone instead of named after a stand-in The rule stack, folder, filters and preset survive quitting Recent folders Cross-filesystem moves fall back to copy-and-delete A command line over the same engine

---

Install

⬇ Download from zsync.eu/zrename

The homepage detects your system and hands you the right file. Every build is
also on GitHub Releases.

Platform	File
Windows 10/11	ZRename_x.y.z_x64-setup.exe — NSIS installer
Debian / Ubuntu	ZRename_x.y.z_amd64.deb
Fedora / RHEL	ZRename-x.y.z-1.x86_64.rpm
Any Linux	ZRename_x.y.z_amd64.AppImage

The Linux packages install a desktop entry, so ZRename appears in your
application menu and in a file manager's Open with for folders.

---

Quick start

Drop a folder onto the window, pick it with the folder button, or name it on the
command line:

zrename-desktop ~/Pictures/import
zrename-desktop ~/Pictures/import --preset "Photos → date-based"

Then stack up rules, read the preview, and press Apply. The button always
states the number — Apply 1,281 renames — and it is disabled while any
collision is unresolved.

Three presets ship with the app:

Preset	What it does
Photos → date-based	IMG_4821.JPG → 2026-08-14_01.jpg, numbered within each folder, from the EXIF capture date
TV episodes → S01E02	Normalises season and episode markers, pads them to two digits, tidies dot-separated names
Sanitise for USB/FAT32	Strips characters Windows rejects, transliterates accents, lowercases extensions

---

The rules

Find & replace — plain or regex

Plain text by default, case-insensitive. Tick Regex for full expressions
with $1 back-references, and the editor grows a live tester showing the match,
the capture groups and the result against the row you have selected.

The replacement may contain metadata tokens. A $ that came out of a token is
escaped automatically, so a filename containing $1 cannot accidentally become
a capture reference.

Numbering — sequential, sorted, per folder

Start, step and zero-padding, ascending or descending, restarting in each folder
or running straight through. The order is chosen explicitly: by name, by
natural order (so img2 comes before img10), by size, or by modification
or creation date.

Numbering is resolved in a sequential pass before the parallel one, so the
result is identical every time however many cores run it.

Sanitise — make a name safe to carry anywhere

Removes characters the target filesystem rejects, so a FAT32 stick is cleaned
more strictly than an ext4 disk. Optionally transliterates Unicode to ASCII
(München → Munchen), collapses runs of spaces, and trims the trailing dots
and spaces Windows silently drops.

Move into folders — turn renaming into organising

A template such as %exif:DateTimeOriginal:%Y%/%exif:DateTimeOriginal:%m% files
each result into 2026/08/. Folders are created as needed, and a leading / or
.. is stripped so results always stay under the folder you loaded.

CSV mapping — names produced somewhere else

Point it at a file of old,new lines and only the listed files are renamed.
Anything missing from the list is left alone.

---

Metadata tokens

Resolved per file, lazily, and cached for the session. A file that has no value
for a token gets a configurable stand-in — or, if you prefer, is left alone
entirely.

Namespace	Examples
exif	%exif:DateTimeOriginal:%Y-%m-%d% · %exif:Model% · %exif:Make% · %exif:Y-m-d%
id3	%id3:artist% · %id3:album% · %id3:title% · %id3:track% · %id3:year%
video	%video:width% · %video:height% · %video:fps% · %video:duration% — needs ffprobe
pdf	%pdf:title% · %pdf:author% · %pdf:pages% · %pdf:creator%
file	%file:stem% · %file:ext% · %file:size% · %file:created:%Y% · %file:modified%
folder	%folder:name% · %folder:parent%
hash	%hash:crc32% · %hash:blake3%
counter	%counter:3% · %index:2%

The format segment is %ns:key:fmt%. Because strftime itself uses %, there is
an unambiguous form when a template gets hairy:

%exif:DateTimeOriginal{%Y-%m-%d}%

---

Presets

Plain TOML in a directory, so a preset can be shared, diffed, version-controlled
and edited by hand.

name = "Photos → date-based"
description = "Names each photo after the day it was taken."

[[rules]]
kind = "template"
template = "%exif:DateTimeOriginal:%Y-%m-%d%"

[[rules]]
kind = "number"
start = 1
pad = 2
reset_per_folder = true
sort = "natural"

[[rules]]
kind = "extension"
mode = "lower"

Ctrl S saves the current stack as one.

---

Filters

Applied before any rule runs, so the preview only ever shows files you meant.

Filter	Example
Extensions	jpg, png, pdf
Only these (glob)	IMG_*, DSC_*
Exclude (glob)	*.tmp, .git/*
Name matches (regex)	^S\d{2}E\d{2}
Size range	1MB to 500MB — also 1.5 GiB, 500k
Modified range	a date on either side
Subfolders	on, with an optional depth limit
Folders and hidden files	off by default

---

Command line

The same engine backs a CLI, so a preset built by clicking behaves identically
in a script or a cron job.

zrename init-presets                                             # write the starter presets
zrename presets                                                  # list what is available
zrename run ~/Pictures/import --preset "Photos → date-based"     # dry run: prints the plan
zrename run ~/Pictures/import --preset "Photos → date-based" --apply
zrename history                                                  # past batches
zrename undo                                                     # put the last one back
zrename undo 2026-09-04T14-06-08.770                             # or a specific one

Flag
--recursive, --max-depth N	walk subfolders
--ext jpg,png	restrict by extension
--csv plan.csv	write the plan instead of printing it
--on-conflict skip|suffix|overwrite	what to do about a taken name
--paranoid	hardlink originals before touching anything
--apply	commit; without it nothing is written

Note

The desktop binary is zrename-desktop and the command-line one is
zrename, so installing both leaves you with two commands that do not
fight over a name.

---

Keyboard

Ctrl K	Command palette	Alt ↑/↓	Reorder the selected rule
Ctrl Enter	Apply	Ctrl D	Duplicate the selected rule
Ctrl Z	Undo the last batch	Ctrl N	Add a rule
Ctrl F	Filter the preview	Ctrl S	Save the stack as a preset
F5	Re-read the folder	Delete	Remove the selected rule
Space	Toggle the selected rule

---

Performance

Measured, not estimated.

Workload	Time
2,000 JPEGs → dated, numbered names (scan, plan, validate)	18 ms
…applied to disk, with the journal flushed first	103 ms
…all 2,000 reverted by a separate process	25 ms
100,000 files scanned, planned and validated	665 ms
100,000 files where every row also attempts an EXIF lookup	902 ms

Rules run under rayon, the preview table
is virtualised, and the plan stays in Rust — the interface asks only for the
window of rows it is currently showing, so a 100,000-file preview never
serialises 100,000 rows.

---

Where things live

	Linux	Windows
Presets	~/.config/zrename/presets/*.toml	%APPDATA%\zrename\presets
Undo journal	~/.local/share/zrename/journal/*.json	%LOCALAPPDATA%\zrename\journal

A journal entry records {from, to, size, mtime, inode} per file, which is what
lets undo tell "this is the file I renamed" from "something else is here
now".

---

Building from source

Needs pnpm, a Rust toolchain, and the
Tauri v2 prerequisites.

git clone https://github.com/TheHolyOneZ/ZRename
cd ZRename
pnpm install

pnpm tauri dev      # run it
pnpm tauri build    # bundle it

Layout

crates/zrename-core/   the engine: scan, plan, validate, journal, execute.
                       No Tauri, no UI, and where most of the tests live.
crates/zrename-cli/    the command line, over the same engine
src-tauri/             the desktop shell and its IPC commands
src/                   React 19 frontend

Tests

cargo test --workspace   # 254: rules, token parsing, filesystem profiles, cycle
                         # detection, journal round-trips, real renames on disk
pnpm test                # 30: diff rendering, size parsing, rule summaries

The Windows job in CI runs the same suite on a windows-latest runner, so NTFS
behaviour is checked against real NTFS rather than only simulated.

---

Licence

GPL-3.0 · built by TheHolyOneZ

zsync.eu/zrename — homepage & downloads


GitHub — source · TheHolyOneZ — the author · zsync.eu — more projects · zlogic.eu — game mods

_{Keywords · bulk rename · batch rename files · mass file renamer · rename multiple files at once · regex file rename · rename photos by EXIF date · rename by metadata · ID3 rename music files · Linux file renamer · Windows bulk rename tool · Advanced Renamer alternative · Bulk Rename Utility alternative · krename alternative · PowerRename alternative · free open source renamer · undo file rename · batch rename with preview · Rust · Tauri}