API Reference - Khanh Docs

Main Function#

`prompt()`#

The main function exported by format-prompt. It's a template tag function that formats template literal strings.

function prompt(  strings: TemplateStringsArray,  ...values: unknown[]): string

Parameters:

strings - The template strings array from a template literal
values - The interpolated values from the template literal

Returns: A formatted string with cleaned whitespace

Example:

import { prompt } from 'format-prompt';const result = prompt`  This is a prompt  with multiple lines`;// Returns: "This is a prompt\nwith multiple lines"

Behavior#

Whitespace Handling#

The prompt function processes whitespace in the following way:

Leading whitespace: Removed from each line
Trailing whitespace: Removed from each line
Multiple spaces: Collapsed to single spaces within lines
Line breaks: Preserved as single line breaks
Final trim: The entire result is trimmed of leading/trailing whitespace

Examples#

Basic Formatting#

const result = prompt`  Hello  World  `;// Returns: "Hello World"

Multi-line Formatting#

const result = prompt`  Line 1  Line 2  Line 3`;// Returns: "Line 1\nLine 2\nLine 3"

Preserving Structure#

const result = prompt`  Title    Body paragraph    Footer`;// Returns: "Title\n\nBody paragraph\n\nFooter"

Interpolation#

The function supports standard template literal interpolation:

const name = 'Alice';const age = 30;const result = prompt`Name: ${name}, Age: ${age}`;// Returns: "Name: Alice, Age: 30"

Interpolated values are converted to strings using String():

const obj = { key: 'value' };const num = 42;const bool = true;const result = prompt`  Object: ${obj}  Number: ${num}  Boolean: ${bool}`;// Object is converted to "[object Object]"// Number is converted to "42"// Boolean is converted to "true"

Type Definitions#

Function Signature#

export function prompt(  strings: TemplateStringsArray,  ...values: unknown[]): string

TemplateStringsArray#

This is a built-in TypeScript type representing the strings in a template literal:

interface TemplateStringsArray extends ReadonlyArray<string> {  readonly raw: ReadonlyArray<string>;}

Use Cases#

AI Prompt Formatting#

function createChatPrompt(systemRole: string, userMessage: string) {  return prompt`    System: ${systemRole}        User: ${userMessage}        Assistant:  `;}

Email Template#

function formatEmail(to: string, subject: string, body: string) {  return prompt`    To: ${to}    Subject: ${subject}        ${body}  `;}

Code Generation#

function generateFunction(name: string, params: string[], body: string) {  return prompt`    function ${name}(${params.join(', ')}) {      ${body}    }  `;}

Performance#

The prompt function is designed to be lightweight and efficient:

Uses native string operations
Minimal overhead for template literal processing
No external dependencies for the core formatting logic
Small bundle size (~1KB minified + gzipped)

Best Practices#

Use for prompts: This function is optimized for formatting AI prompts and similar text
Preserve structure: The function maintains your intended line break structure
Interpolate variables: Use template literal interpolation for dynamic content
Code readability: Write prompts with proper indentation in your code - the function will clean it up

Common Patterns#

Conditional Content#

function createPrompt(includeContext: boolean, context?: string) {  return prompt`    Answer the following question.        ${includeContext && context ? `Context: ${context}\n\n` : ''}        Provide a detailed response.  `;}

List Generation#

function formatList(items: string[]) {  return prompt`    Items:    ${items.map(item => `  * ${item}`).join('\n')}  `;}

Nested Templates#

const section1 = prompt`Section 1 content`;const section2 = prompt`Section 2 content`;const combined = prompt`  ${section1}    ${section2}`;