Bulk rename files

Bulk rename files from one complete preview

A useful batch rename starts before any file bytes are copied. Choose individual files or a folder and the workbench lists each relative path, byte size, and proposed output path. The list is sorted by the exact relative source path, so sequence numbers do not depend on an invisible order chosen by a browser or operating system. Review that order first, then build the rule that turns the old names into the new ones.

The page does not rename your originals in place. It creates new, byte-identical copies for download. That distinction is visible beside the pickers and on the receipt because a browser download and a filesystem rename have different risks. Finder, File Explorer, PowerRename, and desktop batch-renaming applications can change the selected folder itself. This page deliberately leaves that folder untouched and gives you a package you can inspect first.

Find, replace, prefix, suffix, regex, and sequence

Literal find and replace is the safe default. It changes the matching text wherever it occurs in the filename stem, with an explicit case-sensitive switch. Prefix and suffix add stable text around the transformed stem. Regex mode uses a Unicode regular expression for jobs such as moving a captured date or removing a numbered camera prefix. An invalid expression, or an expression that can match an empty string forever, is a planning error rather than a partial batch.

Sequence is optional. Its start, step, minimum padding, separator, and before-or-after position are all visible inputs. A start of 7, step of 3, and padding of 3 produces 007, 010, 013, and so on in the displayed order. Changing any part of the recipe invalidates an older ZIP immediately. The old download cannot remain beside a preview that now says something different.

Extensions are protected unless you include them

By default, the final extension is outside the rename rule. In quarterly.report.PDF, the stem is quarterly.report and the protected extension is .PDF. Its spelling and case remain exactly as selected. A dotfile such as .gitignore has no extension in this model. Turn on Include extension only when the extension is genuinely part of the requested replacement; the preview then shows the complete result before export.

No output name is silently cleaned. Slashes, path traversal, control characters, Windows device names, trailing dots or spaces, and path segments that are too long are shown as errors on the source row that produced them. Fix the rule and the row updates. This keeps the preview authoritative instead of letting the ZIP writer secretly replace a character after you approved the name.

Collisions are blocked, not numbered behind your back

Two names that differ only by letter case can coexist on some disks and overwrite each other on others. Unicode can create the same problem: a visible accented letter may be encoded as one code point or as a base letter plus combining mark. The collision check compares a conservative normalized, case-folded form within each output folder. Both rows become red before packaging if their portable destinations collide.

The tool never resolves that conflict by appending (1). Automatic deduplication would make the downloaded archive disagree with the approved preview. Folder selection keeps valid relative subfolders, but only the filename at the end of each path changes. Folder names are never rewritten, flattened, or moved into a different hierarchy.

The ZIP contains unchanged bytes and a manifest

Export begins only after the plan is valid and at least one name changes. A local worker reads source files in bounded chunks, advances progress from actual bytes, writes the exact previewed paths, and includes JSON and CSV manifests. Each manifest row records source path, target path, byte count, modification time, rule, and the source and output SHA-256 values. The reverse mapping is audit data for the downloaded copies; it is not a claim that this page changed or can roll back your original folder.

The completed archive is reopened in this tab. A file is counted as verified only when its archive path, byte count, and SHA-256 match the selected source. A read failure or verification mismatch names the affected file and withholds a success receipt. This follows the same local package discipline used by the bulk image resizer, while avoiding that tool’s image decode and re-encode because a rename should never alter content.

Large batches use browser memory without an input-size ban

The page does not reject a source file because of its byte size. It shows the shared memory advice before packaging and then attempts the real work. The ZIP writer switches to ZIP64 metadata when classic ZIP fields cannot describe the real entry count, offsets, or byte lengths. It assembles chunked Blob parts rather than demanding one archive-sized typed array. If a constrained tab still runs out of memory, the page reports that browser failure and keeps the selected sources and recipe available for a smaller retry.

Progress names the current phase and counts bytes actually read, packaged, and verified; it is not a timer. Cancel checks between chunks and leaves the inventory and recipe in place. A worker crash returns the interface to an idle, retryable state. Choosing a new source or changing a rule stops the old generation so a stale result can never appear beside newer settings.

Everything stays in this browser tab

Filenames, relative paths, timestamps, manifests, hashes, source bytes, and archive bytes stay on this device. No upload endpoint or third-party renaming service is used. A small local planner starts after file selection or a rule change so even a difficult regex cannot block the page controls. It reads names and metadata only. Hashing, ZIP64 assembly, and read-back code are imported from CreatorValet’s own origin only after the export button, and that is also the first point at which source file bytes are read. The local-processing verification guide explains how to inspect that boundary yourself.

This route changes names, not file contents. To create web-safe text from a headline, use the URL slug generator. To prove one downloaded file still matches a known original outside this receipt, use the SHA-256 checksum tool. AI-generated names, EXIF-driven photo naming, cloud-drive mutation, directory rename, and generated shell scripts are separate jobs and are not hidden modes here.