Literal text replacement pack

Change one exact phrase across a small set of UTF-8 text files. Review each before/after copy, exclude files, then save a ZIP. Originals are never overwritten.

Up to 50 files, 1 MiB each and 5 MiB total. UTF-8 only. Folder paths and filesystem metadata are not retained.

Case-sensitive, non-overlapping literal matches. No regular expressions, whole-word filtering or multiple rules. A blank replacement deletes matches. Both fields are single-line.

No files chosen.

Choose files or load the example.

Inputs remain in page memory. Edits clear the review and disable saving until rebuilt. This is text replacement, not validation or sanitization of the resulting file format.

Finish a small batch edit without replacing originals

Use this tool when the same exact label, address or wording appears in several local plain-text files. Choose the files, enter one phrase to find and its replacement, then review the before and after text for each file. Only matched files selected in the review are included as edited copies in the ZIP. Keep the originals until you have checked the output in its intended application.

For example, the fictional files contain Old Studio in an English note and a Tamil note. The example changes that phrase to New Studio. A third file has no match and is reported without an edited copy. If a matching file should stay unchanged, untick it before saving. This makes a small, deliberate edit set rather than silently changing every chosen file.

Exactly what matches

The search is case-sensitive and literal. A dot, dollar sign, bracket or other punctuation is ordinary text, not a regular expression or replacement instruction. Matches do not overlap: finding aa in aaa makes one replacement. Matching is by the exact Unicode sequence; visually similar characters or differently normalized accented text are not treated as equal. Matches can occur inside longer words, markup, comments or quoted data. Review the context before using the edited file.

Find must contain 1–1,000 characters and replacement may contain 0–1,000. Both are single-line inputs; line breaks and control characters other than tabs are rejected. This tool does not normalize line endings. It cannot replace a phrase that spans lines, apply a list of rules or parse a programming language.

UTF-8 and output boundaries

Every chosen file must decode as valid UTF-8. Invalid byte sequences, NUL characters and other non-text control characters stop the whole review. This is a limited text check, not a general binary-file detector. Do not use Word documents, spreadsheets, PDFs, images or other packaged formats. A file can be valid UTF-8 and still be unsuitable for the intended change.

The output retains an existing UTF-8 byte-order mark and the original CRLF, CR or LF line endings outside replaced text. It does not preserve original timestamps, permissions, folders, extended attributes or a digital signature. Replacement can invalidate JSON, HTML, code, checksums or signed content. The app does not run, render or certify the edited file.

Review, selection and the saved pack

The preview shows full before and after text for files up to 20,000 characters. Larger files show their first 20,000 characters and are explicitly labelled as truncated; inspect the downloaded copy in an editor before using it. The match count covers the full file. Untick any changed file that you do not want to include. A no-match file and a replacement identical to the find text produce no edited copy.

The ZIP contains selected edited copies under numbered safe filenames in edited/, plus manifest.json and README.txt. The manifest maps each original filename to its output path or exclusion reason, with match counts, input and output byte sizes, UTF-8 BOM presence and SHA-256 digests. The manifest records your find and replacement text, so it may itself contain private information. Hashes help compare bytes; they do not prove identity, authorization or file safety. ZIP timestamps are fixed, and generated paths intentionally differ from original names to avoid unsafe paths and collisions.

Temporary inputs and limits

Choose at most 50 files, each no larger than 1 MiB and together no larger than 5 MiB. Expanded output is limited to 10 MiB in total. These bounds keep the browser review manageable. Selecting another batch replaces the current list. Changing either phrase or choosing new files clears the previous preview. Clear or reload removes the page state. Downloaded files remain on your device.

The app makes no uploads, hidden requests, tracking calls or browser-storage history. It does not save a workspace, edit files in place, send the ZIP or connect to a repository. File contents are displayed as plain text; markup is never executed. You decide which copies to distribute and should check their contents first.

Reference: Visual Studio Code’s editing documentation describes an editor-based alternative for searching and reviewing replacements across a folder.