@tact-ddd/ids

Type-safe, prefixed, and globally unique identifiers for Domain-Driven Design

You can think of this as: "one global friendly alphabet + a tiny factory that makes branded, prefixed ID types."

This package provides:

• nanoid with a custom alphabet (no confusing chars like 0/O or 1/l)
• short prefixes (e.g. ws_, usr_, ord_, etc.)
• branded string types so TypeScript won't let you mix WorkspaceId and UserId
• validation and type guards for runtime safety

Inspired by Unkey's ID implementation.

Installation

npm install @tact-ddd/ids
# or
bun add @tact-ddd/ids
# or
pnpm add @tact-ddd/ids

Quick Start

import { defineIdType } from '@tact-ddd/ids';

// Define a User ID type
export const UserIdFactory = defineIdType({ prefix: 'usr' });
export type UserId = ReturnType<(typeof UserIdFactory)['parse']>;

// Create a new ID
const userId = UserIdFactory.create();
// => "usr_A3fK8mN9pQrS2tUv"

// Use it in your domain
function findUser(id: UserId) {
	// ...
}

findUser(userId); // ✅ OK

Core Concepts

1. Brand Helper

The package uses TypeScript's nominal typing pattern to create distinct types at compile-time while keeping them as strings at runtime:

export type Brand<TBase, TBrand extends string> = TBase & {
	readonly __brand: TBrand;
};

This ensures that UserId and WorkspaceId are treated as different types by TypeScript, even though they're both strings under the hood.

2. Friendly Alphabet

The package uses a Base58-like alphabet without ambiguous characters:

const FRIENDLY_ALPHABET = '123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz';

This excludes:

• 0 (zero) and O (capital o)
• 1 (one), l (lowercase L), and I (capital i)

3. Type-Safe Factory

The defineIdType function creates a factory for each ID type with:

• A branded type (e.g., WorkspaceId)
• A create() function that generates prefix_xxx strings
• A parse() function that validates and brands existing strings
• An is() type guard for runtime checks

API Reference

defineIdType<P>(config: IdConfig<P>)

Creates a type-safe ID factory with the specified configuration.

Parameters

interface IdConfig<P extends string> {
	prefix: P; // The prefix for this ID type (e.g., "usr", "ws")
	length?: number; // Optional: override default length (default: 16)
	strict?: boolean; // Optional: enforce alphabet validation (default: false)
}

Returns

An object with the following methods and properties:

{
  prefix: string;           // The prefix used
  length: number;           // The core ID length (excluding prefix)
  create: () => Id;         // Generate a new ID
  parse: (raw: string) => Id;  // Validate and brand an existing string
  is: (raw: string) => raw is Id;  // Type guard for runtime checks
}

Usage Examples

Basic Usage

import { defineIdType } from '@tact-ddd/ids';

// Define your ID types
export const WorkspaceIdFactory = defineIdType({ prefix: 'ws' });
export type WorkspaceId = ReturnType<(typeof WorkspaceIdFactory)['parse']>;

export const UserIdFactory = defineIdType({ prefix: 'usr' });
export type UserId = ReturnType<(typeof UserIdFactory)['parse']>;

// Create new IDs
const workspaceId = WorkspaceIdFactory.create();
// => "ws_A3fK8mN9pQrS2tUv"

const userId = UserIdFactory.create();
// => "usr_X7yZ4bC6dE9fG2hJ"

Type Safety

TypeScript prevents you from mixing different ID types:

function loadWorkspace(id: WorkspaceId) {
	// ...
}

function findUser(id: UserId) {
	// ...
}

const workspaceId = WorkspaceIdFactory.create();
const userId = UserIdFactory.create();

loadWorkspace(workspaceId); // ✅ OK
findUser(userId); // ✅ OK

// These cause compile-time errors:
loadWorkspace(userId); // ❌ Type error!
findUser(workspaceId); // ❌ Type error!

// Plain strings don't work either:
const plainString = 'ws_abc123';
loadWorkspace(plainString); // ❌ Type error!

Parsing Existing IDs

Use parse() to validate and brand IDs from external sources (e.g., databases, APIs):

const OrderIdFactory = defineIdType({ prefix: 'ord' });
export type OrderId = ReturnType<(typeof OrderIdFactory)['parse']>;

// From a database or API
const rawId = 'ord_K8mN9pQrS2tU';

try {
	const orderId = OrderIdFactory.parse(rawId);
	// orderId is now branded as OrderId
	processOrder(orderId);
} catch (error) {
	console.error('Invalid order ID:', error.message);
}

// Invalid examples that throw errors:
OrderIdFactory.parse('usr_abc123'); // ❌ Wrong prefix
OrderIdFactory.parse('ord_123'); // ❌ Wrong length
OrderIdFactory.parse('ord_'); // ❌ Empty core

Custom Length

Different ID types can have different lengths:

// Short IDs for frequently used entities
const SessionIdFactory = defineIdType({
	prefix: 'sess',
	length: 12,
});
// => "sess_A3fK8mN9pQr"

// Longer IDs for permanent entities
const DocumentIdFactory = defineIdType({
	prefix: 'doc',
	length: 24,
});
// => "doc_A3fK8mN9pQrS2tUvW4xY7zA2"

Strict Mode

Enable strict mode to validate that IDs only contain characters from the friendly alphabet:

const ProductIdFactory = defineIdType({
	prefix: 'prod',
	length: 16,
	strict: true,
});

// Valid - uses only friendly alphabet characters
ProductIdFactory.parse('prod_A3fK8mN9pQrS2tUv'); // ✅ OK

// Invalid - contains disallowed characters
ProductIdFactory.parse('prod_0O1lI_invalid!!'); // ❌ Throws error

Without strict mode (default), parse() only checks prefix and length.

Runtime Type Guards

Use the is() method for runtime type checks:

const WorkspaceIdFactory = defineIdType({ prefix: 'ws' });

function handleId(raw: string) {
	if (WorkspaceIdFactory.is(raw)) {
		// TypeScript knows 'raw' is WorkspaceId here
		loadWorkspace(raw);
	} else {
		console.log('Not a valid workspace ID');
	}
}

Advanced Patterns

Central Prefix Registry

Create a single source of truth for all your ID prefixes:

// id-registry.ts
import { defineIdType } from '@tact-ddd/ids';

export const prefixes = {
	workspace: 'ws',
	user: 'usr',
	team: 'team',
	project: 'proj',
	task: 'task',
	comment: 'cmt',
	api_key: 'key',
} as const;

export type IdKind = keyof typeof prefixes;

// Define all ID factories
export const WorkspaceIdFactory = defineIdType({ prefix: prefixes.workspace });
export type WorkspaceId = ReturnType<(typeof WorkspaceIdFactory)['parse']>;

export const UserIdFactory = defineIdType({ prefix: prefixes.user });
export type UserId = ReturnType<(typeof UserIdFactory)['parse']>;

export const TeamIdFactory = defineIdType({ prefix: prefixes.team });
export type TeamId = ReturnType<(typeof TeamIdFactory)['parse']>;

// ... export other factories

// Optional: Generic ID creator
export function createId(kind: IdKind) {
	switch (kind) {
		case 'workspace':
			return WorkspaceIdFactory.create();
		case 'user':
			return UserIdFactory.create();
		case 'team':
			return TeamIdFactory.create();
		case 'project':
			return ProjectIdFactory.create();
		case 'task':
			return TaskIdFactory.create();
		case 'comment':
			return CommentIdFactory.create();
		case 'api_key':
			return ApiKeyIdFactory.create();
	}
}

Domain Entity Integration

Use branded IDs in your domain entities:

import { WorkspaceId, UserId } from './id-registry';

interface Workspace {
	id: WorkspaceId;
	name: string;
	ownerId: UserId;
	createdAt: Date;
}

interface User {
	id: UserId;
	email: string;
	workspaceIds: WorkspaceId[];
}

// Repository with type-safe IDs
class WorkspaceRepository {
	async findById(id: WorkspaceId): Promise<Workspace | null> {
		// Implementation
	}

	async save(workspace: Workspace): Promise<void> {
		// Implementation
	}
}

Database Integration

When working with databases, parse IDs on the way in and out:

import { WorkspaceIdFactory, WorkspaceId } from './id-registry';

// Reading from database
async function getWorkspaceFromDb(rawId: string): Promise<Workspace> {
	const id = WorkspaceIdFactory.parse(rawId); // Validate and brand

	const row = await db.query('SELECT * FROM workspaces WHERE id = $1', [id]);

	return {
		id,
		name: row.name,
		// ...
	};
}

// Writing to database
async function saveWorkspace(workspace: Workspace): Promise<void> {
	await db.query(
		'INSERT INTO workspaces (id, name) VALUES ($1, $2)',
		[workspace.id, workspace.name] // ID is already a string at runtime
	);
}

API Route Handlers

Validate IDs from request parameters:

import { WorkspaceIdFactory } from './id-registry';

app.get('/workspaces/:id', async (req, res) => {
	try {
		const workspaceId = WorkspaceIdFactory.parse(req.params.id);
		const workspace = await workspaceRepo.findById(workspaceId);

		if (!workspace) {
			return res.status(404).json({ error: 'Workspace not found' });
		}

		res.json(workspace);
	} catch (error) {
		res.status(400).json({ error: 'Invalid workspace ID' });
	}
});

Testing

The package includes comprehensive tests. Run them with:

bun test

Example test patterns:

import { describe, expect, test } from 'bun:test';
import { defineIdType } from '@tact-ddd/ids';

describe('ID Generation', () => {
	test('creates valid IDs with correct format', () => {
		const UserId = defineIdType({ prefix: 'user', length: 10 });
		const id = UserId.create();

		expect(typeof id).toBe('string');
		expect(id.startsWith('user_')).toBe(true);
		expect(id.length).toBe('user_'.length + 10);
	});

	test('parses valid IDs and rejects invalid ones', () => {
		const OrderId = defineIdType({ prefix: 'order', length: 12 });
		const validId = OrderId.create();

		expect(OrderId.parse(validId)).toBe(validId);
		expect(() => OrderId.parse('invalid_123')).toThrow();
		expect(() => OrderId.parse('order_123')).toThrow(); // Too short
	});
});

Best Practices

1. Export both the factory and the type:

   export const UserIdFactory = defineIdType({ prefix: 'usr' });
   export type UserId = ReturnType<(typeof UserIdFactory)['parse']>;

2. Use descriptive prefixes: Keep them short (2-4 chars) but meaningful

   • ✅ usr, ws, proj, ord
   • ❌ u, w, p, o (too cryptic)
   • ❌ user, workspace (too long)

3. Parse IDs at boundaries: Always parse IDs from external sources (APIs, databases) before using them in your domain logic

4. Use strict mode for user-facing IDs: Enable strict: true when parsing IDs that users might manually enter

5. Document your prefix registry: Keep all your prefixes in one place with documentation

Why Use This Package?

Type Safety

• Compile-time checks prevent mixing different ID types
• Auto-completion in your IDE for ID-related operations
• Refactoring confidence - TypeScript will catch breaking changes

Runtime Safety

• Prefix validation ensures IDs are used in the correct context
• Length validation catches truncated or malformed IDs
• Alphabet validation (strict mode) ensures data integrity

Developer Experience

• Clear error messages when validation fails
• Consistent API across all ID types
• Easy testing with predictable behavior

Production Ready

• No ambiguous characters reduces user errors
• Globally unique with nanoid's collision resistance
• Compact - shorter than UUIDs while remaining unique
• Fast - nanoid is highly optimized

License

MIT

Credits

• Inspired by Unkey's ID implementation
• Built on nanoid by Andrey Sitnik