Реклама мешает? Идти Без рекламы Сегодня 

Генератор CUID2

ДанныеРазработчикСлучайныйБезопасность

Реклама · УДАЛИТЬ?

ВХОД

ВЫХОД

Клиентская сторона

Generated CUID2 IDs

CUID v1 IDs (for comparison)

ID Structure Breakdown

Отпечаток

Browser Fingerprint

About CUID2

Реклама · УДАЛИТЬ?

Гид

Генератор CUID2

Generate cryptographically secure CUID2 IDs entirely in your browser. CUID2 is a modern, collision-resistant identifier format that combines a high-entropy random salt, monotonic counter, timestamp, and browser fingerprint, then runs the result through SHA3-512 to produce a base36 string. Every ID starts with a letter, so values stay safe to use in databases, URLs, and most identifier columns.

Как использовать

Pick an ID length between 8 and 32 characters. The default of 24 is a good fit for most apps.
Choose how many IDs to generate at once (up to 100).
Optionally enable the CUID v1 comparison to see the legacy format side by side.
Нажмите Создайте to produce IDs, then copy or download the results.
Inspect the structure breakdown to see exactly how each piece (timestamp, counter, salt, fingerprint) feeds into the final ID.

Возможности

Настраиваемая длина – 8 to 32 characters, with collision-space estimates updated live.
Bulk generation – up to 100 IDs per click for seeding databases or testing.
Database-safe first letter – every ID begins with a–z so it works in identifier columns and CSS selectors.
SHA3-512 hashing – matches the official CUID2 spec for the entropy mixing step.
Fingerprint exposure – view the browser fingerprint that contributes to each generated ID.
Structure breakdown – see the timestamp, counter, salt, and fingerprint pieces that combined to make a sample ID.
CUID v1 comparison – generate matching CUID v1 IDs side by side to understand the format differences.
Cryptographic randomness – использует crypto.getRandomValues() rather than predictable PRNGs.
100% клиентская сторона – nothing is sent to a server, IDs never leave your browser.

 Часто задаваемые вопросы

What makes CUID2 different from UUID v4?

UUID v4 uses 122 bits of pure randomness with no structural protections against bad PRNGs or duplicated environments. CUID2 deliberately mixes a timestamp, monotonic counter, host fingerprint, and a cryptographic random salt before hashing them with SHA3-512. The result is host-independent, hard to forge, and uses only lowercase letters and digits, making CUID2 IDs URL-safe, database-safe, and resilient to weak randomness on a single machine.
Why do CUID2 IDs always start with a letter?

Many systems treat identifiers that start with a digit as numbers (CSS selectors, GraphQL field names, JavaScript variable names, some database identifier columns). By forcing the first character to be a–z, CUID2 IDs can be used as identifiers in nearly any environment without escaping or quoting. The remaining characters come from the base36-encoded SHA3 hash and may be letters or digits.
Are CUID2 IDs sortable by creation time?

No. CUID v1 was loosely sortable because the timestamp appeared in plain text near the front of the ID. CUID2 hashes the timestamp together with the salt and fingerprint, so the resulting characters reveal nothing about creation order. If you need time-ordered IDs, use a format such as ULID, KSUID, or Snowflake. If you want unguessable, opaque IDs that are safe to expose externally, CUID2 is a strong fit.
How likely is a CUID2 collision?

At the default 24-character length the ID space is roughly 26 × 36²³, which is far beyond any realistic generation rate. The birthday-paradox 50% collision threshold lands well above quintillions of IDs. Increasing the length to 32 makes collisions effectively impossible at any realistic scale, while shorter lengths (8–10) are appropriate only for low-volume use cases like share codes or short URLs.
What is the role of the browser fingerprint?

The fingerprint is a hash derived from environment data such as user agent, screen size, timezone, and locale, mixed with a fresh random salt. It ensures that two different machines generating IDs at the same millisecond produce different hash inputs. The fingerprint never identifies you personally — it is one piece of entropy among several feeding into the SHA3 hash and is not stored or transmitted.