jlevy/repren

repren

---

[!TIP] Coding agents: tell your agent to run uvx repren@latest --help and follow the instructions to install repren as a skill. Once installed, it uses repren automatically for bulk refactors and renames. See Agent Use.

[!NOTE] Zero runtime dependencies. repren is a single file that imports only the Python standard library, so there is no transitive dependency tree to audit or be compromised: the only code uvx repren@latest runs is repren itself.

---

Rename Anything

repren is a powerful CLI string replacement and file renaming tool for use by agents or humans for almost any search-and-replace or renaming task.

It is small, self-contained, self-documenting, and works on Python 3.10-3.14 with zero runtime dependencies. Essentially, it is a general-purpose, brute-force text file refactoring tool.

Coding agents: point your agent at repren and it installs itself as a skill:

Run uvx repren@latest --help and follow the instructions to install repren as a skill.

repren’s --help ends with the exact install commands. Once installed, the agent uses repren automatically for bulk refactors and renames. (See Agent Use.)

Zero runtime dependencies. Installing repren (or running uvx repren@latest) pulls in nothing but repren itself and the Python standard library, so there is no transitive dependency tree to audit or be compromised. It is a single file you can read end to end before trusting it. The dev-only tooling used to build and test repren (pytest, ruff, and so on) is never installed when you use it.

For example, repren could rename occurrences of certain names in a set of source files, while simultaneously renaming the files and directories according to the same pattern and handling all case variations.

It’s more powerful than classic options like perl -pie, rpl, or sed:

• Replacements: It allows rewriting file contents according to one or more literal or regular expression patterns.

• Renames: It can also apply the patterns to rename or move files according to replacements on their full paths, creating directories as needed.

• Regexes: It supports fully expressive regular expressions substitutions, including matching groups for back substitutions (like \1, \2, etc.).

• Simultaneous renames: It performs simultaneous renamings: you can make as many replacements as you want and you can rename “foo” as “bar”, and “bar” as “foo” at once, without requiring a temporary intermediate rename.

• Good hygiene: It is careful: it has a nondestructive “dry run” mode and prints clear stats on its changes. It leaves backups. File operations are done atomically, so interruptions never leave a previously existing file truncated or partly edited.

• Case preserving options: It supports “magic” case-preserving renames that let you find and rename identifiers with case variants (lowerCamel, UpperCamel, lower_underscore, and UPPER_UNDERSCORE) consistently.

• Dry run, backups, and undo: It has convenient options for dry run, undo (restoring backups), and cleanup (deleting backups).

• Text or JSON output: It supports human-readable text output (default) or machine-parseable JSON output (--format=json) for easy integration with scripts and agents.

• Self-documenting: It is packaged with its own nice documentation! Run repren --docs for full documentation.

If file paths are provided, repren replaces those files in place, leaving a backup with extension “.orig” (controlled by the --backup-suffix option).

If directory paths are provided, it applies replacements recursively to all files in the supplied paths that are not in the exclude pattern. If no arguments are supplied, it reads from stdin and writes to stdout.

Examples

Patterns can be supplied in a text file, with one or more replacements consisting of regular expression and replacement. For example:

# Sample pattern file
frobinator<tab>glurp
WhizzleStick<tab>AcmeExtrudedPlasticFunProvider
figure ([0-9+])<tab>Figure \1

(Where <tab> is an actual tab character.) Each line is a replacement. Empty lines and #-prefixed comments are ignored.

As a short-cut, a single replacement can be specified on the command line using --from (match) and --to (replacement).

Examples:

# Here `patfile` is a patterns file.
# Rewrite stdin:
repren --patterns=patfile < input > output

# Shortcut with a single pattern replacement (replace foo with bar):
repren --from=foo --to=bar < input > output

# Rewrite a few files in place, also requiring matches be on word breaks:
repren --patterns=patfile --word-breaks myfile1 myfile2 myfile3

# Rewrite whole directory trees. Since this is a big operation, we use
# `-n` to do a dry run that only prints what would be done:
repren -n --patterns=patfile --word-breaks --full mydir1

# Now actually do it:
repren --patterns=patfile --word-breaks --full mydir1

# Same as above, for all case variants:
repren --patterns=patfile --word-breaks --preserve-case --full mydir1

# Same as above but including only .py files and excluding the tests directory
# and any files or directories starting with test_:
repren --patterns=patfile --word-breaks --preserve-case --full --include='.*[.]py$' --exclude='tests|test_.*' mydir1

Usage

Run repren --docs for full usage and flags.

If file paths are provided, repren replaces those files in place, leaving a backup with extension “.orig”. If directory paths are provided, it applies replacements recursively to all files in the supplied paths that are not in the exclude pattern. If no arguments are supplied, it reads from stdin and writes to stdout.

Comparison to Alternatives

There are many tools for search/replace and refactoring. Here’s how repren compares:

“Automatic backups” means a backup copy is written by default; repren writes .orig files automatically, while sed/awk/perl and rnr can only with an explicit flag.

“Review capabilities” covers previewing and reversing changes: repren and rnr offer a dry run and undo, sd prints a preview, and fastmod, ast-grep, and comby step through changes with interactive terminal prompts (on by default in fastmod, behind a flag in ast-grep and comby).

“Atomic file operations” means edits are written to a temp file and renamed into place, so an interrupted run never leaves a half-written file. repren and sd do this; rnr only renames; the classic tools vary; and fastmod, ast-grep, and comby overwrite in place.

When to use each:

• repren: The default for renaming and search/replace across files: content edits, file and directory renames, case-variant refactors, and simultaneous swaps, all with dry-run, backups, and undo.
• sed/awk/perl and rnr: For renaming or search-and-replace, prefer repren. It is safer and more capable for almost any such task (dry-run, backups and undo, simultaneous swaps, case-variant refactors, and combined content-and-path renames). Keep sed/awk/perl for what they are built for, such as streaming transforms, field extraction, and small inline programs. rnr only renames files, which repren already does.
• sd: A fast, simple find/replace when you do not need file renaming, case preservation, or multi-pattern swaps.
• fastmod: Interactive, human-reviewed replacements.
• ast-grep: Language-aware, structural refactoring where you match code structure (for example, function calls) rather than text. It also ships an official agent skill and MCP server, so it pairs well with repren for AST-aware edits.
• comby: Structural matching across languages (for example, balanced braces) without learning full AST syntax.

Installation

repren has zero runtime dependencies (just Python 3.10+), so installing it never adds a transitive dependency tree to your environment. It’s easiest to install with uv:

# Install as a tool:
uv tool install repren

# Or run directly without installing:
uvx repren@latest --help

Or, since it’s just one file, you can copy the repren.py script somewhere convenient and make it executable.

A note on freshness and safety. These examples use @latest for simplicity. Because repren has zero runtime dependencies, the only code fetched and run is repren itself. There is no transitive dependency tree to audit or be compromised, so @latest carries far less supply-chain risk here than for a typical package. If you want an extra safety margin, you can opt into uv’s release “cool-off” (which excludes very recent uploads), or pin an exact version:

# Skip uploads newer than 14 days, then run the latest within that window:
UV_EXCLUDE_NEWER="14 days" uvx repren@latest --help

# Or pin an exact version:
uvx [email protected] --help

Agent Use

Point your agent at repren’s self-documenting help to install the skill:

Run uvx repren@latest --help and follow the instructions to install repren as a skill.

repren’s --help ends with the exact install commands. (uvx runs repren with no prior install; see the note on @latest and safety under Installation.)

Install the skill yourself (the same thing the agent does):

# Install into the current project (run from the repo; shareable via git):
uvx repren@latest --install-skill --project

# Or install globally, for every project:
uvx repren@latest --install-skill --global

Re-run any time to update an existing install.

How it works

repren is self-documenting: repren --help and repren --docs are the source of truth for every flag, and --format=json gives agents machine-parseable output. The skill is just a thin pointer to those commands, so it never drifts out of date.

Installing the skill writes one SKILL.md to both discovery locations, so every agent finds it:

• .agents/skills/repren/: the portable, cross-agent location (Codex, Gemini, pi, …)
• .claude/skills/repren/: the Claude Code mirror

repren is a general-purpose utility with no per-project config, so the skill invokes it through the zero-install runner uvx repren@latest. There’s no need to add repren as a project dependency. Because repren has zero runtime dependencies, the only code a runner ever fetches and runs is repren itself; for an extra safety margin you can opt into uv’s release cool-off (UV_EXCLUDE_NEWER, see Installation) in your environment, which applies to the skill’s invocations too.

Scope is resolved like git config: implicit when unambiguous, a clear error when not, so a stray --install-skill never silently rewrites your global agent surfaces:

• Inside a git repo, --project is implied (so uvx repren --install-skill just works).
• In an ambiguous spot (your home directory, or outside any repo) repren refuses and tells you to pass --project (optionally with --dir DIR or --no-repo-check) or --global.

Manual install: run uvx repren --skill and save the output to .agents/skills/repren/SKILL.md (and/or .claude/skills/repren/SKILL.md), under ~ for a global install or the project root for a project install.

Learn more: Claude Code docs and Skills repository.

Try It

Let’s try a simple replacement in my working directory (which has a few random source files):

$ repren --from=frobinator-server --to=glurp-server --full --dry-run .
Dry run: No files will be changed
Using 1 patterns:
  'frobinator-server' -> 'glurp-server'
Found 102 files in: .
- modify: ./site.yml: 1 matches
- rename: ./roles/frobinator-server/defaults/main.yml -> ./roles/glurp-server/defaults/main.yml
- rename: ./roles/frobinator-server/files/deploy-frobinator-server.sh -> ./roles/glurp-server/files/deploy-frobinator-server.sh
- rename: ./roles/frobinator-server/files/install-graphviz.sh -> ./roles/glurp-server/files/install-graphviz.sh
- rename: ./roles/frobinator-server/files/frobinator-purge-old-deployments -> ./roles/glurp-server/files/frobinator-purge-old-deployments
- rename: ./roles/frobinator-server/handlers/main.yml -> ./roles/glurp-server/handlers/main.yml
- rename: ./roles/frobinator-server/tasks/main.yml -> ./roles/glurp-server/tasks/main.yml
- rename: ./roles/frobinator-server/templates/frobinator-webservice.conf.j2 -> ./roles/glurp-server/templates/frobinator-webservice.conf.j2
- rename: ./roles/frobinator-server/templates/frobinator-webui.conf.j2 -> ./roles/glurp-server/templates/frobinator-webui.conf.j2
Read 102 files (190382 chars), found 2 matches (0 skipped due to overlaps)
Dry run: Would have changed 2 files, including 0 renames

That was a dry run, so if it looks good, it’s easy to repeat that a second time, dropping the --dry-run flag. If this is in git, we’d do a git diff to verify, test, then commit it all. If we messed up, there are still .orig files present.

Patterns

Patterns can be supplied using the --from and --to syntax above, but that only works for a single pattern.

In general, you can perform multiple simultaneous replacements by putting them in a patterns file. Each line consists of a regular expression and replacement. For example:

# Sample pattern file
frobinator<tab>glurp
WhizzleStick<tab>AcmeExtrudedPlasticFunProvider
figure ([0-9+])<tab>Figure \1

(Where <tab> is an actual tab character.)

Empty lines and #-prefixed comments are ignored. Capturing groups and back substitutions (such as \1 above) are supported.

Examples

# Here `patfile` is a patterns file.
# Rewrite stdin:
repren --patterns=patfile < input > output

# Shortcut with a single pattern replacement (replace foo with bar):
repren --from=foo --to=bar < input > output

# Rewrite a few files in place, also requiring matches be on word breaks:
repren --patterns=patfile --word-breaks myfile1 myfile2 myfile3

# Rewrite whole directory trees. Since this is a big operation, we use
# `-n` to do a dry run that only prints what would be done:
repren -n --patterns=patfile --word-breaks --full mydir1

# Now actually do it:
repren --patterns=patfile --word-breaks --full mydir1

# Same as above, for all case variants:
repren --patterns=patfile --word-breaks --preserve-case --full mydir1

Backup Management

Repren provides tools for managing backup files created during operations:

Undo Changes

If you need to revert changes, use --undo with the same patterns as the original operation:

# Original operation:
repren --from=OldClass --to=NewClass --full src/

# Undo the changes:
repren --undo --from=OldClass --to=NewClass --full src/

The undo command:

• Finds all .orig backup files
• Uses the patterns to determine which files were renamed
• Restores the original files and removes renamed files
• Skips with warnings if timestamps look wrong or files are missing

Clean Backups

When you’re satisfied with your changes, remove backup files:

# Remove all .orig backup files:
repren --clean-backups src/

# Dry run to see what would be removed:
repren --clean-backups --dry-run src/

# Remove backups with custom suffix:
repren --clean-backups --backup-suffix=.bak src/

Complete Workflow

A typical workflow:

# 1. Preview changes
repren --dry-run --from=foo --to=bar --full mydir/

# 2. Execute changes (creates .orig backups)
repren --from=foo --to=bar --full mydir/

# 3. Review and test your changes

# 4. Either undo if something went wrong:
repren --undo --from=foo --to=bar --full mydir/

# 4. Or clean up backups when satisfied:
repren --clean-backups mydir/

Notes

• All pattern matching is via standard Python regular expressions.

• As with sed, replacements are made line by line by default. Memory permitting, replacements may be done on entire files using --at-once.

• As with sed, replacement text may include backreferences to groups within the regular expression, using the usual syntax: \1, \2, etc.

• In the pattern file, both the regular expression and the replacement may contain the usual escapes \\n, \\t, etc. (To match a multi-line pattern, containing \\n, you must use --at-once.)

• Replacements are all matched on each input file, then all replaced, so it’s possible to swap or otherwise change names in ways that would require multiple steps if done one replacement at a time.

• If two patterns have matches that overlap, only one replacement is applied, with preference to the pattern appearing first in the patterns file.

• If one pattern is a subset of another, consider if --word-breaks will help.

• If patterns have special characters, --literal may help.

• The case-preserving option works by adding all case variants to the pattern replacements, e.g. if the pattern file has foo_bar -> xxx_yyy, the replacements fooBar -> xxxYyy, FooBar -> XxxYyy, FOO_BAR -> XXX_YYY are also made. Assumes each pattern has one casing convention.

• The same logic applies to filenames, with patterns applied to the full file path with slashes replaced and then parent directories created as needed, e.g. my/path/to/filename can be rewritten to my/other/path/to/otherfile. (Use caution and test with -n, especially when using absolute path arguments!)

• Files are never clobbered by renames. If a target already exists, or multiple files are renamed to the same target, numeric suffixes will be added to make the files distinct (".1", “.2”, etc.).

• Files are created at a temporary location, then renamed, so original files are left intact in case of unexpected errors. File permissions are preserved.

• Backups are created of all modified files, with the suffix “.orig”. The suffix can be customized with --backup-suffix.

• By default, recursive searching omits paths starting with “.” (including .git/). This may be adjusted with --exclude. Files ending in the backup suffix (.orig by default) are always ignored. repren does not read .gitignore: any other paths you want to skip (node_modules/, build/, dist/, vendored code) must be named explicitly with --exclude, and richer scoping is done with --include/--exclude (preview with --dry-run).

• Data is handled as bytes internally, allowing it to work with any encoding or binary files. File contents are not decoded unless necessary (e.g., for logging). However, patterns are specified as strings in the pattern file and command line arguments, and file paths are handled as strings for filesystem operations.

Contributing

Contributions and issues welcome! Check the output of the test script and if it has changed or needs updating, and commit the clean log changes if you submit a PR.

License

MIT.