fast-strip-ansi

A performance-optimized ANSI escape sequence stripper, built on top of the vt-push-parser crate.

Full VT-100/ANSI support.

No startup cost for regex compilation.

Usage

use fast_strip_ansi::*;
let input = b"Hello, world!\x1b[31m\nHello, world!\x1b[0m";
assert_eq!(strip_ansi_bytes(input), b"Hello, world!\nHello, world!".as_slice());
let input = "Hello, world!\x1b[31m\nHello, world!\x1b[0m";
assert_eq!(strip_ansi_string(input), "Hello, world!\nHello, world!");

Security and Correctness

fast-strip-ansi is correct and secure. It contains a true VT-100/ANSI state machine which handles corner cases not handled by other crates.

Performance

fast-strip-ansi the fastest and most correct library for stripping ANSI escape sequences.

Note that fast-strip-ansi has wider support for ANSI escape sequences than most crates but is still the fastest option.

It is significantly faster than strip-ansi and strip-ansi-escapes in all modes.

xychart
    title "Performance"
    x-axis [fast-strip-ansi, fast-strip-ansi-callback, strip-ansi, strip-ansi-escapes]
    y-axis "Time (in µs)" 0 --> 100
    bar [22.12, 11.26, 39.05, 292.5]
    bar [11.89, 6.841, 19.21, 277.4]

%% Your markdown viewer may not render the chart correctly.
%% View in a mermaid chart viewer like <https://mermaid.live>

Raw performance data:

from cargo bench on an M3 MacBook Pro

0 suffix means no ansi content, 100 means significant ansi content

Crate versions: