You are viewing the docs for v1.3, which is not the latest release.
View latest

Which Languages Have Dedicated Kestra Script Plugins

For the complete documentation index, see llms.txt. For a full content snapshot, see llms-full.txt. Append .md to any kestra.io/docs/* URL for plain Markdown.

Kestra lets you run code in many languages inside your workflows.

Choose the right way to run a language

Use a dedicated script plugin when Kestra provides one for your language. Use the Shell plugin when you need to run another language or compile code inside a container.

Languages with dedicated plugins

Kestra provides dedicated script plugins for these languages:

Each of these plugins provides two task types:

  • Script for short inline code in your flow definition.
  • Commands for code stored in files or split across multiple commands.

Here is a minimal example that uses the Python Script task:

yamlid: python_script_example
namespace: company.team

tasks:
  - id: run_python
    type: io.kestra.plugin.scripts.python.Script
    containerImage: python:3.12-slim
    script: |
      print("Hello from Python")

Use the Commands task when you want to run a file from namespaceFiles or execute multiple commands in sequence:

yamlid: python_commands_example
namespace: company.team

tasks:
  - id: run_python_file
    type: io.kestra.plugin.scripts.python.Commands
    namespaceFiles:
      enabled: true
    containerImage: python:3.12-slim
    commands:
      - python hello.py

Language-specific guides

Use these guides for complete examples, outputs, metrics, and dependency management:

Run other languages with the Shell plugin

Use io.kestra.plugin.scripts.shell.Commands when your language does not have a dedicated plugin or when you want to compile and run code in a container.

This approach works best when:

  • the language runtime is available in the container image
  • your commands can be executed from a shell
  • you want to use namespaceFiles or inputFiles for source code

For outputs and metrics, use the same ::{}:: syntax documented for Shell tasks. See Shell outputs and metrics.

Rust example

This example compiles and runs a Rust file stored in namespaceFiles:

yamlid: rust_example
namespace: company.team

tasks:
  - id: run_rust
    type: io.kestra.plugin.scripts.shell.Commands
    taskRunner:
      type: io.kestra.plugin.scripts.runner.docker.Docker
    containerImage: rust:1.82
    namespaceFiles:
      enabled: true
    commands:
      - rustc hello_world.rs
      - ./hello_world

The hello_world.rs file can contain:

rustfn main() {
    println!("Hello, World!");
}

See the full Rust guide for outputs and file handling.

Java example

If you need to run Java code without building a custom plugin, you can compile and run it from a container:

yamlid: java_example
namespace: company.team

tasks:
  - id: run_java
    type: io.kestra.plugin.scripts.shell.Commands
    taskRunner:
      type: io.kestra.plugin.scripts.runner.docker.Docker
    containerImage: eclipse-temurin:21
    namespaceFiles:
      enabled: true
    commands:
      - javac HelloWorld.java
      - java HelloWorld
javaclass HelloWorld {
    public static void main(String[] args) {
        System.out.println("Hello, World!");
    }
}

TypeScript example

You can run TypeScript with the Node.js plugin by compiling it to JavaScript first:

yamlid: typescript_example
namespace: company.team

tasks:
  - id: run_typescript
    type: io.kestra.plugin.scripts.node.Commands
    namespaceFiles:
      enabled: true
    containerImage: node:22-slim
    commands:
      - npm install --save-dev typescript
      - npx tsc example.ts
      - node example.js

The example.ts file can contain:

typescripttype User = {
  name: string;
  age: number;
};

function isAdult(user: User): boolean {
  return user.age >= 18;
}

const justine: User = {
  name: "Justine",
  age: 23,
};

console.log(isAdult(justine));

For more background, see the official Node.js with TypeScript guide.

Use a custom Docker image for extra dependencies

Use a custom Docker image when you need tools or libraries that are not present in the default runtime image.

The Dockerfile below adds Go to a Kestra image:

dockerfileFROM kestra/kestra:latest
USER root

RUN apt-get update -y && apt-get install -y wget && \
    wget -qO- https://go.dev/dl/go1.24.3.linux-amd64.tar.gz | tar -C /usr/local -xzf - && \
    echo 'export PATH=$PATH:/usr/local/go/bin' > /etc/profile.d/golang.sh

ENV PATH="/usr/local/go/bin:${PATH}"

Point your docker-compose.yml file to that Dockerfile:

yamlservices:
  kestra:
    build:
      context: .
      dockerfile: Dockerfile
    image: kestra-go:latest

After you start Kestra with docker compose up -d, you can run Go code with the Process task runner:

yamlid: golang_process
namespace: company.team

tasks:
  - id: go_custom_dependencies
    type: io.kestra.plugin.scripts.go.Script
    taskRunner:
      type: io.kestra.plugin.core.runner.Process
    beforeCommands:
      - go mod init go_script
      - go get github.com/go-gota/gota/dataframe
      - go mod tidy
    script: |
      package main

      import (
          "os"
          "github.com/go-gota/gota/dataframe"
          "github.com/go-gota/gota/series"
      )

      func main() {
          names := series.New([]string{"Alice", "Bob", "Charlie"}, series.String, "Name")
          ages := series.New([]int{25, 30, 35}, series.Int, "Age")
          df := dataframe.New(names, ages)
          file, _ := os.Create("output.csv")
          defer file.Close()
          df.WriteCSV(file)
      }
    outputFiles:
      - output.csv