Introducing pikaid: Small · Sortable · Secure Unique IDs

Hey everyone!

I’m Lara, a Belgian web developer, and this is the very first repo I’ve kicked off entirely on my own. I wanted a compact, drop-in unique ID you can stash in any database and have sort itself—no extra tricks needed. UUIDs felt bulky, ULIDs a bit low on randomness, so I thought: why not stick to 26 chars, switch to Base36, and pack in 96 bits of cryptographic entropy? That’s Pikaid in a nutshell.

🎯 The Problem

• UUIDs are 36 characters with dashes, and you need extra logic to sort them.
• ULIDs nail the 26-char size and sorting, but cap out at 80 bits of randomness.
• We’re often forced to pick between length, sortability, or entropy.

Putting those elements together in the mixer, I created a single ID that’s:

1. 26 chars long
2. Sortable by creation time out of the box
3. Packed with 96 bits of randomness
4. Simple: just 0–9a–z, no extra symbols

🔧 Meet Pikaid

• Length: 26 chars
• Alphabet: lowercase Base36 (0–9, a–z)
• Structure: 7 chars timestamp (seconds since epoch, Base36, zero-padded), 19 chars randomness (96 bits, Base36, zero-padded)
• Sortable: lex order = creation time
• Secure: ~7.9×10²⁸ combos → collisions are practically impossible

⚡ Quickstart (PHP)

<?php

use Pikaid\Pikaid;

require 'vendor/autoload.php';

$id = Pikaid::generate();

echo $id; // e.g. 'kfz1w5x0ab9c23d4e5f6g7h8ij'

JavaScript version coming soon—stay tuned!

🚀 Roadmap & Contributions

I’m working on:

• Full JS implementation
• ???

This is my first public repo, so I probably messed up a few things. Please:

• Flag any unclear spots in the spec
• Suggest extra use cases or edge cases
• Open issues & PRs with ideas
• Guide me on open-source best practices

Your feedback will help me level up and make Pikaid rock-solid.

🔗 Get Involved

👉 Specifications Repo: GitHub →
👉 PHP Repo: GitHub →
👉 License: MIT

Big thanks for giving Pikaid a spin! I hope you find it useful—drop your thoughts, point out any mistakes, and let’s build something awesome together.