Programming language: Go
License: MIT License
Latest version: v0.17.1

script alternatives and similar packages

Based on the "DevOps Tools" category.
Alternatively, view script alternatives based on common mentions on social networks and blogs.

Do you think we are missing an alternative of script or a related project?

Add another 'DevOps Tools' Package


Go ReferenceGo Report CardMentioned in Awesome GoCircleCI

import "github.com/bitfield/script"

[Magical gopher logo](img/magic.png)

What is script?

script is a Go library for doing the kind of tasks that shell scripts are good at: reading files, executing subprocesses, counting lines, matching strings, and so on.

Why shouldn't it be as easy to write system administration programs in Go as it is in a typical shell? script aims to make it just that easy.

Shell scripts often compose a sequence of operations on a stream of data (a pipeline). This is how script works, too.

This is one absolutely superb API design. Taking inspiration from shell pipes and turning it into a Go library with syntax this clean is really impressive.\ —Simon Willison

Read more: Scripting with Go

Quick start: Unix equivalents

If you're already familiar with shell scripting and the Unix toolset, here is a rough guide to the equivalent script operation for each listed Unix command.

Unix / shell script equivalent
(any program name) Exec
[ -f FILE ] IfExists
> WriteFile
>> AppendFile
$* Args
basename Basename
cat File / Concat
curl Do / Get / Post
cut Column
dirname Dirname
echo Echo
find FindFiles
grep Match / MatchRegexp
grep -v Reject / RejectRegexp
head First
jq JQ
ls ListFiles
sed Replace / ReplaceRegexp
sha256sum SHA256Sum / SHA256Sums
tail Last
uniq -c Freq
wc -l CountLines
xargs ExecForEach

Some examples

Let's see some simple examples. Suppose you want to read the contents of a file as a string:

contents, err := script.File("test.txt").String()

That looks straightforward enough, but suppose you now want to count the lines in that file.

numLines, err := script.File("test.txt").CountLines()

For something a bit more challenging, let's try counting the number of lines in the file that match the string Error:

numErrors, err := script.File("test.txt").Match("Error").CountLines()

But what if, instead of reading a specific file, we want to simply pipe input into this program, and have it output only matching lines (like grep)?


Just for fun, let's filter all the results through some arbitrary Go function:


That was almost too easy! So let's pass in a list of files on the command line, and have our program read them all in sequence and output the matching lines:


Maybe we're only interested in the first 10 matches. No problem:


What's that? You want to append that output to a file instead of printing it to the terminal? You've got some attitude, mister. But okay:


We're not limited to getting data only from files or standard input. We can get it from HTTP requests too:

// Output:
// London: 🌦   +13°C

That's great for simple GET requests, but suppose we want to send some data in the body of a POST request, for example. Here's how that works:


If we need to customise the HTTP behaviour in some way, such as using our own HTTP client, we can do that:

    Timeout: 10 * time.Second,

Or maybe we need to set some custom header on the request. No problem. We can just create the request in the usual way, and set it up however we want. Then we pass it to Do, which will actually perform the request:

req, err := http.NewRequest(http.MethodGet, "http://example.com", nil)
req.Header.Add("Authorization", "Bearer "+token)

The HTTP server could return some non-okay response, though; for example, “404 Not Found”. So what happens then?

In general, when any pipe stage (such as Do) encounters an error, it produces no output to subsequent stages. And script treats HTTP response status codes outside the range 200-299 as errors. So the answer for the previous example is that we just won't see any output from this program if the server returns an error response.

Instead, the pipe “remembers” any error that occurs, and we can retrieve it later by calling its Error method, or by using a sink method such as String, which returns an error value along with the result.

Stdout also returns an error, plus the number of bytes successfully written (which we don't care about for this particular case). So we can check that error, which is always a good idea in Go:

_, err := script.Do(req).Stdout()
if err != nil {

If, as is common, the data we get from an HTTP request is in JSON format, we can use JQ queries to interrogate it:

data, err := script.Do(req).JQ(".[0] | {message: .commit.message, name: .commit.committer.name}").String()

We can also run external programs and get their output:


Note that Exec runs the command concurrently: it doesn't wait for the command to complete before returning any output. That's good, because this ping command will run forever (or until we get bored).

Instead, when we read from the pipe using Stdout, we see each line of output as it's produced:

PING ( 56 data bytes
64 bytes from icmp_seq=0 ttl=64 time=0.056 ms
64 bytes from icmp_seq=1 ttl=64 time=0.054 ms

In the ping example, we knew the exact arguments we wanted to send the command, and we just needed to run it once. But what if we don't know the arguments yet? We might get them from the user, for example.

We might like to be able to run the external command repeatedly, each time passing it the next line of data from the pipe as an argument. No worries:

script.Args().ExecForEach("ping -c 1 {{.}}").Stdout()

That {{.}} is standard Go template syntax; it'll substitute each line of data from the pipe into the command line before it's executed. You can write as fancy a Go template expression as you want here (but this simple example probably covers most use cases).

If there isn't a built-in operation that does what we want, we can just write our own, using Filter:

script.Echo("hello world").Filter(func (r io.Reader, w io.Writer) error {
    n, err := io.Copy(w, r)
    fmt.Fprintf(w, "\nfiltered %d bytes\n", n)
    return err
// Output:
// hello world
// filtered 11 bytes

The func we supply to Filter takes just two parameters: a reader to read from, and a writer to write to. The reader reads the previous stages of the pipe, as you might expect, and anything written to the writer goes to the next stage of the pipe.

If our func returns some error, then, just as with the Do example, the pipe's error status is set, and subsequent stages become a no-op.

Filters run concurrently, so the pipeline can start producing output before the input has been fully read, as it did in the ping example. In fact, most built-in pipe methods, including Exec, are implemented using Filter.

If we want to scan input line by line, we could do that with a Filter function that creates a bufio.Scanner on its input, but we don't need to:

script.Echo("a\nb\nc").FilterScan(func(line string, w io.Writer) {
    fmt.Fprintf(w, "scanned line: %q\n", line)
// Output:
// scanned line: "a"
// scanned line: "b"
// scanned line: "c"

And there's more. Much more. Read the docs for full details, and more examples.

A realistic use case

Let's use script to write a program that system administrators might actually need. One thing I often find myself doing is counting the most frequent visitors to a website over a given period of time. Given an Apache log in the Common Log Format like this: - - [30/Jun/2019:17:06:15 +0000] "GET / HTTP/1.1" 200 2028 "https://example.com/ "Mozilla/5.0 (Linux; Android 8.0.0; FIG-LX1 Build/HUAWEIFIG-LX1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/64.0.3282.156 Mobile Safari/537.36"

we would like to extract the visitor's IP address (the first column in the logfile), and count the number of times this IP address occurs in the file. Finally, we might like to list the top 10 visitors by frequency. In a shell script we might do something like:

cut -d' ' -f 1 access.log |sort |uniq -c |sort -rn |head

There's a lot going on there, and it's pleasing to find that the equivalent script program is quite brief:

package main

import (

func main() {

Let's try it out with some [sample data](testdata/access.log):



See pkg.go.dev for the full documentation, or read on for a summary.


These are functions that create a pipe with a given contents:

Source Contents
Args command-line arguments
Do HTTP response
Echo a string
Exec command output
File file contents
FindFiles recursive file listing
Get HTTP response
IfExists do something only if some file exists
ListFiles file listing (including wildcards)
Post HTTP response
Slice slice elements, one per line
Stdin standard input


Filters are methods on an existing pipe that also return a pipe, allowing you to chain filters indefinitely. The filters modify each line of their input according to the following rules:

Filter Results
Basename removes leading path components from each line, leaving only the filename
Column Nth column of input
Concat contents of multiple files
Dirname removes filename from each line, leaving only leading path components
Do response to supplied HTTP request
Echo all input replaced by given string
Exec filtered through external command
ExecForEach execute given command template for each line of input
Filter user-supplied function filtering a reader to a writer
FilterLine user-supplied function filtering each line to a string
FilterScan user-supplied function filtering each line to a writer
First first N lines of input
Freq frequency count of unique input lines, most frequent first
Get response to HTTP GET on supplied URL
Join replace all newlines with spaces
JQ result of jq query
Last last N lines of input
Match lines matching given string
MatchRegexp lines matching given regexp
Post response to HTTP POST on supplied URL
Reject lines not matching given string
RejectRegexp lines not matching given regexp
Replace matching text replaced with given string
ReplaceRegexp matching text replaced with given string
SHA256Sums SHA-256 hashes of each listed file

Note that filters run concurrently, rather than producing nothing until each stage has fully read its input. This is convenient for executing long-running comands, for example. If you do need to wait for the pipeline to complete, call Wait.


Sinks are methods that return some data from a pipe, ending the pipeline and extracting its full contents in a specified way:

Sink Destination Results
AppendFile appended to file, creating if it doesn't exist bytes written, error
Bytes data as []byte, error
CountLines number of lines, error
Read given []byte bytes read, error
SHA256Sum SHA-256 hash, error
Slice data as []string, error
Stdout standard output bytes written, error
String data as string, error
Wait none
WriteFile specified file, truncating if it exists bytes written, error

What's new

Version New
v0.21.0 HTTP support: Do, Get, Post
v0.20.0 JQ


See the [contributor's guide](CONTRIBUTING.md) for some helpful tips if you'd like to contribute to the script project.


Gopher image by MariaLetta