Functions

Functions in BAML define type-safe contracts between your application and AI models.

Syntax

function FunctionName(param: Type) -> ReturnType {
    client ModelSpec
    prompt #"
        Template content
    "#
}

Components

Component	Description	Example
name	Function identifier (must start with capital letter)	ExtractEmail
parameters	Typed input parameters	text: string, data: MyClass
return_type	Guaranteed output type	string, MyClass, "a" | "b"
client	LLM specification	"openai/gpt-4o", GPT4
prompt	Jinja template	Block string with {{ }}

Basic Example

function ExtractEmail(text: string) -> string {
    client "openai/gpt-4o"
    prompt #"
        Extract the email address from the following text:
        {{ text }}

        {{ ctx.output_format }}
    "#
}

Client Specification

Shorthand Syntax

function MakeHaiku(topic: string) -> string {
    client "openai/gpt-4o"
    prompt #"
        Write a haiku about {{ topic }}.
    "#
}

Named Client

client<llm> MyClient {
    provider "openai"
    options {
        model "gpt-4o"
    }
}

function MakeHaiku(topic: string) -> string {
    client MyClient
    prompt #"
        Write a haiku about {{ topic }}.
    "#
}

Type System

Functions support all BAML types:

Built-in Types

Type	Description
string	Text data
int	Integer numbers
float	Decimal numbers
bool	True/false values
Type[]	Arrays
map<K, V>	Key-value pairs
"a" | "b"	Literal unions

Custom Types

class Person {
    name string
    age int
    contacts Contact[]
}

class Contact {
    type "email" | "phone"
    value string
}

function ParsePerson(data: string) -> Person {
    client "openai/gpt-4o"
    prompt #"
        {{ ctx.output_format }}

        {{ _.role('user') }}
        {{ data }}
    "#
}

Prompt Templates

Jinja Syntax

prompt #"
    Input data: {{ input_data }}

    {% if condition %}
        Conditional content
    {% endif %}

    {{ ctx.output_format }}
"#

Special Variables

Variable	Description
ctx.output_format	Auto-generated format instructions based on return type
ctx.client	Selected client and model name
_.role()	Define message role (system, user, assistant)

Error Handling

BAML functions automatically handle:

• JSON parsing errors (auto-corrected)
• Type mismatches (detected and reported)
• Network and rate limit errors (propagated to caller)

Generated Client Usage

Python

from baml_client import b
from baml_client.types import Person

async def process() -> Person:
    result = await b.ParsePerson("John Doe, 30 years old...")
    print(result.name)  # Type-safe access
    return result

TypeScript

import { b } from 'baml_client';
import { Person } from 'baml_client/types';

async function process(): Promise<Person> {
    const result = await b.ParsePerson("John Doe, 30 years old...");
    console.log(result.name);  // Type-safe access
    return result;
}

See Also

• Prompts
• Tests
• Client Configuration
• Prompt Syntax