Starlark Built-in Modules

Minimal's Lark steps run Starlark — a sandboxed, deterministic subset of Python. Built-in modules extend what you can do beyond core Starlark.

Load a module at the top of your code block before using any function from it:

load("math", "upby10")
load("strings", "slugify")

---

math

load("math", "upby10")

Function	Description	Example
upby10(n)	Multiplies n by 10	upby10(5) → 50

---

strings

load("strings", "slugify")

Function	Description	Example
slugify(s)	Converts string to URL-safe slug	slugify("Hello World") → "hello-world"

---

Core Starlark — No Load Required

These are available in every Lark step without loading:

Function	Description	Example
len(x)	Length of list or string	len(step_0) → 42
str(x)	Convert to string	str(123) → "123"
int(x)	Convert to integer	int("42") → 42
range(n)	Integer range	range(3) → [0, 1, 2]
print(x)	Log to server output	print(step_0)

---

What Starlark Does Not Support

Starlark is intentionally limited for safety and determinism. These Python features are not available:

Not supported	Use instead
import	Use load()
f-strings	String concatenation — "hello " + name
try / except	Validate input before using it
class	Use dicts instead
Generators	Use explicit for loops with append
None	Use "" or 0 as empty values

tip

If you need something not available in Starlark, do the heavy transformation in SQL and use Lark only for shaping the output.