How to Generate Random UUIDs (Free Online UUID Generator)

Every record in a database needs a unique identifier. Every API resource needs a stable reference. Every distributed system needs IDs that do not collide even when generated by different machines simultaneously. UUIDs solve all three problems with a single, standardized format.

This guide explains what UUIDs are, how they work, when to use them, and how to generate them instantly online.

What is a UUID?

A UUID (Universally Unique Identifier) is a 128-bit identifier formatted as a sequence of hexadecimal characters:

550e8400-e29b-41d4-a716-446655440000

The format is always 32 hexadecimal characters in five groups, separated by hyphens: 8-4-4-4-12 characters. This produces a string that is:

• Unique: The probability of two randomly generated UUIDs being identical is astronomically low
• Stateless: Generated without coordination between systems — no central registry, no database query
• Portable: Works as a string in any programming language, database, URL, or log file

UUIDs are sometimes called GUIDs (Globally Unique Identifiers) — the same format, just Microsoft's name for it.

To generate a UUID immediately, use the FixTools UUID Generator.

UUID versions

There are several UUID versions, each generated differently:

UUID v4 (random)

The most widely used version. 122 bits of random data, 6 bits reserved for version and variant. Every call produces a completely random value with no relationship to previous UUIDs, the generating machine, or the timestamp.

f47ac10b-58cc-4372-a567-0e02b2c3d479

Use v4 for: user IDs, session tokens, record identifiers, API keys.

UUID v1 (time-based)

Encodes the current timestamp (in 100-nanosecond intervals since October 1582) and the MAC address of the generating machine. Unique across time and machines, but leaks information about when and where it was generated.

Use v1 for: systems where you need to recover the generation timestamp from the ID itself.

UUID v5 (name-based, SHA-1)

Deterministic: given the same namespace and name inputs, always produces the same UUID. Used when you want a UUID for a specific resource (a URL, a username) that stays consistent.

Use v5 for: content-addressable identifiers, stable IDs for known entities.

UUID v7 (time-ordered, newer standard)

A newer format that is sortable by generation time (like v1) but does not expose the MAC address (unlike v1). Ideal for database primary keys because time-ordered UUIDs append to B-tree indexes efficiently rather than scattering randomly.

Use v7 for: database primary keys where both uniqueness and sort order matter.

How to generate a random UUID

Online: Use the FixTools UUID Generator. Click once to generate a UUID, or specify a quantity to generate multiple at once — useful for seeding test data or pre-generating IDs for a batch operation.

In code:

JavaScript (browser and Node.js 14.17+):

const id = crypto.randomUUID();
// Returns something like: '3b12f1df-5232-4804-897e-917bf397618a'

Python:

import uuid
id = str(uuid.uuid4())
# '3b12f1df-5232-4804-897e-917bf397618a'

Node.js (with the uuid package):

import { v4 as uuidv4 } from 'uuid';
const id = uuidv4();

Go:

import "github.com/google/uuid"
id := uuid.New().String()

SQL (PostgreSQL has a built-in function):

SELECT gen_random_uuid();
-- f47ac10b-58cc-4372-a567-0e02b2c3d479

When to use UUIDs

Database primary keys

The traditional alternative is an auto-increment integer (1, 2, 3...). Auto-increment is simple and index-friendly, but has limitations:

• The next ID requires a database round-trip to obtain
• IDs are sequential, which reveals information (a user with ID 3 tells you this is an early user)
• Merging data from multiple databases produces collisions

UUIDs eliminate all three problems. The application can generate an ID before inserting a record, multiple databases can merge without ID conflicts, and IDs reveal nothing about the record's position in the dataset.

Distributed systems

When multiple services, servers, or clients generate IDs independently, they need a strategy that guarantees uniqueness without coordination. UUID v4's statistical uniqueness makes coordination unnecessary — each service generates IDs locally with no risk of collision.

Session and token identifiers

Login sessions, password reset tokens, email verification links, and OAuth tokens all benefit from being UUID-formatted: long enough to be unguessable, short enough to fit in a database column or URL, and standardized enough to work across any language or framework.

Referencing resources in APIs

REST API endpoints commonly use UUIDs to identify resources:

GET /api/users/f47ac10b-58cc-4372-a567-0e02b2c3d479
GET /api/documents/550e8400-e29b-41d4-a716-446655440000

This avoids exposing sequential integer IDs that reveal business information (total user count, order volume) and prevents users from guessing adjacent IDs.

Test data and fixtures

When seeding a development database, a staging environment, or unit test fixtures, you need IDs that are unique but do not conflict with each other or with production data. Generating a batch of UUIDs is faster than manually assigning sequential IDs and prevents accidental collisions if environments share infrastructure.

UUID vs. other ID formats

Format	Example	Unique?	Sortable?	Short?
UUID v4	f47ac10b-58cc-...	Yes (random)	No	No (36 chars)
UUID v7	018e5b09-...	Yes	Yes (time-ordered)	No
ULID	01ARZ3NDEKTSV4RRFFQ69G5FAV	Yes	Yes	Similar
Snowflake ID	1541815603606036480	Yes (distributed)	Yes	Similar
Auto-increment	42	Within one DB	Yes	Yes
NanoID	V1StGXR8_Z5jdHi6B-myT	Yes (random)	No	Shorter

UUID v4 is the default choice when you need universally unique IDs and do not need sort order. Use UUID v7 or ULID when you need time-ordered IDs for database performance.

Common UUID mistakes

Storing UUIDs as strings in a database

A UUID string is 36 characters (32 hex + 4 hyphens). Most databases have a native UUID type (PostgreSQL UUID, MySQL BINARY(16)) that stores the 16 raw bytes instead — half the storage and faster index lookups. Using VARCHAR(36) for UUIDs wastes space and slows queries on large tables.

Using UUID v1 without understanding the privacy implications

UUID v1 encodes the MAC address of the generating machine. If you generate UUID v1 in a client application and include it in logs or API calls, you are leaking the client's hardware identifier. Use v4 for any ID generated on user devices.

Validating UUIDs with a regex that is too loose

A UUID is a specific format. The regex [0-9a-f-]{36} will match it but also match many invalid strings. The correct pattern enforces the group lengths and the version/variant bits:

^[0-9a-f]{8}-[0-9a-f]{4}-4[0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$

This validates UUID v4 specifically (the 4 in the third group and [89ab] at the start of the fourth group).

Generate a UUID now

Use the FixTools UUID Generator to generate one UUID or a batch of UUIDs instantly. Copy individual values or download a list — useful for seeding databases, creating test fixtures, or pre-generating IDs for an import. For broader dummy data including names, emails, and numbers, the Dummy Data Generator generates structured test records you can use directly.