grex — A Command-Line Tool That Generates Regular Expressions from Examples

What it is

grex is a command-line tool written in Rust that generates regular expressions from user-provided example strings. Instead of manually crafting a regex pattern, you give grex a list of strings you want to match, and it produces the simplest regex that covers all examples.

It targets developers, data engineers, and anyone who writes regex infrequently enough that the syntax feels unfamiliar each time. Instead of consulting regex cheat sheets, you describe what you want by example.

How it saves time or tokens

Writing correct regex by hand requires recalling syntax for character classes, quantifiers, anchors, and lookaheads. Testing and debugging takes additional iterations. grex shortens this to a single command. Provide examples, get a working regex. The tool also supports options to produce more or less specific patterns, giving you control over generalization.

How to use

1. Install grex: brew install grex (macOS), cargo install grex (from source), or download a binary.
2. Run grex with your example strings as arguments.
3. grex outputs a regex that matches all provided examples.

Example

# Generate regex for date patterns
grex '2026-01-15' '2025-12-31' '2024-06-01'
# Output: ^\d{4}-\d{2}-\d{2}$

# Generate regex for email-like patterns
grex 'user@example.com' 'admin@test.org' 'info@company.net'
# Output: ^[a-z]+@[a-z]+\.[a-z]{2,3}$

# Use flags for more specific patterns
grex --repetitions '192.168.1.1' '10.0.0.1' '172.16.0.1'

Related on TokRepo

• Automation tools — Tools that speed up repetitive developer tasks
• Coding tools — Development utilities and helpers

Common pitfalls

• grex generates a regex that matches all provided examples, but it may also match strings you did not intend. Provide enough diverse examples to constrain the pattern.
• The generated regex may be more general or more specific than desired. Use flags like --repetitions, --digits, and --words to control the level of generalization.
• grex works best for structural patterns (dates, IPs, codes). For semantic matching (e.g., 'all valid email addresses'), you still need hand-tuned regex or a dedicated parser.