babashka

Babashka

Babashka (bb) is a native Clojure interpreter for scripting with fast startup. Use it when the task looks like shell scripting, project automation, CI glue, small CLIs, file/process automation, or quick data transformation where JVM Clojure startup would be too heavy.

Quick Decision: What do you need?

Task	Go to
Write an executable script, parse args, read stdin, print output	scripts.md
Add or fix bb.edn tasks, task dependencies, shell, clojure, run	tasks.md
Use built-in namespaces, add deps/paths/classpath, load pods	deps-and-pods.md
Work with files, processes, HTTP, JSON/EDN, shell commands, portability	interop.md
Something differs from JVM Clojure, is slow, cannot load, or behaves oddly	anti-patterns.md

Core Mental Model

Babashka runs Clojure through SCI, not JVM bytecode compilation. It starts quickly and ships as a self-contained native binary with a curated set of Clojure namespaces, useful libraries, and selected Java classes. It can interpret compatible Clojure source from the classpath, but it cannot load arbitrary Java classes at runtime.

Choose the form based on audience:

• One-off expression: bb -e '(println (+ 1 2 3))'
• Pipeline transform: bb -i -o '(map clojure.string/upper-case *input*)'
• Executable script: #!/usr/bin/env bb
• Project automation: bb.edn with :tasks
• External capability not built in: dependency, classpath, CLI subprocess, or pod

Useful defaults:

• *command-line-args* contains script or task arguments.
• *in*, *out*, and *err* work like Clojure.
• babashka.fs is the preferred cross-platform filesystem API.
• babashka.process is the preferred subprocess API for non-trivial process work.
• Use clojure.tools.cli for option parsing when scripts grow beyond positional args.