Speed up Claude Code with modern CLI tools

Claude Code runs shell commands constantly — finding files, searching code, processing JSON, running scripts. The default Unix tools (find, grep, sed, node) work, but modern replacements written in Rust and Zig are significantly faster.

Installing these tools and telling Claude Code to prefer them can noticeably speed up your workflow.

A real example

I searched this blog’s codebase for the word “astro”:

• grep: 1.7 seconds
• ripgrep: 0.025 seconds (68x faster)

A typical Claude Code session might run 50-100 searches. At 1.7 seconds each, that’s nearly 3 minutes of waiting. With ripgrep, it’s under 3 seconds total. The difference compounds fast.

The tools

Install on macOS

# Install bun (has its own installer)
curl -fsSL https://bun.sh/install | bash

# Install the rest via Homebrew
brew install ripgrep fd sd jq parallel

Install on Windows

Windows has three package managers. Pick one approach and stick with it.

Option A: winget (built into Windows 11)

winget install Oven-sh.Bun
winget install BurntSushi.ripgrep.MSVC
winget install sharkdp.fd
winget install jqlang.jq

Note: sd is not available via winget. Install it with scoop (scoop install sd) or cargo (cargo install sd).

Option B: scoop (popular with developers, no admin rights needed)

First install scoop if you don’t have it:

Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser
Invoke-RestMethod -Uri https://get.scoop.sh | Invoke-Expression

Then install the tools:

scoop install bun ripgrep fd sd jq

GNU parallel on Windows

GNU parallel is a Unix tool with no native Windows version. Use it through WSL:

wsl --install

Then inside your WSL terminal:

sudo apt-get install parallel

Tell Claude Code to use them

The key step: add these tools to your ~/.claude/CLAUDE.md file so Claude Code knows they’re available. Without this, Claude Code will fall back to the slower defaults.

Add this section to your CLAUDE.md:

## Installed CLI tools

- **bun** is installed — prefer over `node`/`npm` for running scripts, installing packages, and executing JS/TS
- **ripgrep** (`rg`) is installed — prefer over `grep` for shell searches
- **fd** is installed — prefer over `find` for file finding by name/pattern
- **sd** is installed — prefer over `sed` for find-and-replace in files
- **jq** is installed — use for JSON processing in shell pipelines
- **GNU parallel** is installed — use for concurrent shell tasks when beneficial

Why this works

Claude Code reads CLAUDE.md at the start of every session. When it sees these instructions, it will reach for the faster tool every time it needs to search, find, or process something in the shell.

The speed gains compound quickly. A typical coding session involves dozens of file searches, content greps, and script executions. Shaving milliseconds off each one adds up.