Command line and power tools

fd, the modern find

Updated 2026-07-31 · 4 min read

fd is what find would look like designed today: sane defaults, readable syntax, speed.

Install

brew install fd

The defaults that matter

fd report

That alone: searches the current tree, case-insensitively (smart case), skipping hidden files and everything in .gitignore, with coloured output. The find equivalent is a line of flags.

The daily commands

fd invoice ~/Documents          # search a path
fd -e pdf                       # by extension
fd -e pdf invoice               # both
fd -H config                    # include hidden files
fd -I node_modules              # ignore .gitignore rules
fd -t d projects                # directories only
fd -t f -S +100m                # files over 100 MB
fd --changed-within 2d          # modified in the last 2 days

Acting on results

fd -e log -x rm {}              # delete each match
fd -e jpg -X mogrify -resize 50%   # one command, all matches
fd -e md -x wc -l {} \;

-x runs per-file in parallel; -X batches all matches into one invocation.

Regex when you need it

Patterns are regex by default:

fd '^IMG_\d{4}\.(jpg|heic)$'

Glob mode with -g when that reads better: fd -g '*.tar.*'.

Where find still wins

Scripts that must run on any machine (find is always present), exotic predicates (-newer, permission tests), and -prune logic ported from elsewhere. For interactive use, fd wins on ergonomics so thoroughly that find becomes the special-occasion tool.

Common questions

Is fd faster than find?

Usually much faster in practice — parallel traversal plus skipping ignored/hidden trees by default.

Does fd search hidden files?

Not by default; add -H for hidden and -I to ignore .gitignore exclusions.