v1.2.0 - Latest Release

arrow

A fast, composable CLI for modern workflows.

arrow is a command-line toolkit that lets you define, chain, and run complex development pipelines with a single, fast binary. Think of it as a programmable task runner - no DSL, no YAML lock-in, just pure composability from the terminal.

arrow-session
$ arrow init my-project
Created project "my-project"
arrow.yaml config
src/ source
pipelines/ pipeline definitions
 
$ arrow run build
(1/3) lint ........ done
(2/3) test ........ done
(3/3) build ....... done
Completed in 1.24s

Key Features

Blazing Fast
Written in Rust. Parallel pipeline execution with intelligent caching - skip steps that haven't changed.
Composable
Chain commands with pipes, conditionals, and variables. Each step feeds into the next.
Scriptable
Define pipelines as YAML or TOML. Use in CI, locally, or as part of larger scripts.
Cross-Platform
One binary for macOS, Linux, and Windows. No runtime dependencies required.

Why arrow?

Modern development involves many steps - linting, testing, building, deploying. Each tool has its own interface and configuration. arrow wraps them into reusable, composable pipelines that you define once and run everywhere.

The key insight: a pipeline is just a directed graph of commands. arrow resolves dependencies, runs steps in parallel where possible, and caches results so rebuilds are instant. No more waiting for CI to tell you what `npm test` already caught locally.

yaml
# arrow.yaml
version: "1"
pipelines:
  build:
    steps:
      - name: lint
        run: npm run lint
      - name: test
        run: npm run test
        needs: [lint]
      - name: build
        run: npm run build
        needs: [test]

arrow runs build, which resolves the dependency graph and executes lint, then test, then build - all with automatic caching and clear output.

Quick Start

bash
# Install arrow
curl -fsSL https://arrow.sh/install | sh

# Verify
arrow --version

# Create your first project
arrow init demo
cd demo

# Run a pipeline
arrow run

Ready to dive in? Check out Getting Started.