ripgrep (rg) — Recursively Search Directories with Regex

What ripgrep Does

• Recursive search — walks directory tree
• Gitignore aware — skips .gitignore, .ignore, global ignore
• Smart case — --smart-case for case-insensitive when all lower
• File types — --type js or --type-add custom
• Globs — -g include/exclude patterns
• Regex — full Rust regex engine (PCRE2 with -P)
• Context — -A/-B/-C lines around matches
• Output — text, JSON, vimgrep, colored
• Parallelism — multi-threaded by default

Architecture

Single Rust binary. Uses memchr + SIMD for byte-level search, regex crate for patterns. Walks filesystem in parallel (one thread per CPU). Respects .gitignore via the ignore crate.

Self-Hosting

CLI tool, self-contained binary.

Key Features

• Fastest grep alternative (sometimes 10x faster than grep)
• Gitignore/ignore respect
• Parallel search
• JSON output mode
• File type filters
• Unicode aware
• PCRE2 support (with -P)
• Replace mode (-r)
• Integrates with VS Code, Helix, Zed, IntelliJ

Comparison

FAQ

Q: How much faster is it than grep? A: Typically 2-10x. The gap is even bigger on large repositories (Linux kernel scale) because ripgrep runs in parallel and skips .gitignore.

Q: Can it search hidden files? A: They're skipped by default. Enable with -. or --hidden.

Q: Why is it public-domain licensed? A: It's the author BurntSushi's choice. Unlicense is equivalent to waiving copyright — completely free to use.

Sources & Credits

• GitHub: https://github.com/BurntSushi/ripgrep
• License: Unlicense/MIT