Slugkit

SlugKit started as a side project, a playful rebellion against the lifelessness of UUIDs and the tyranny of auto-increment IDs. Now it’s a full-fledged slug generation engine — designed to make identity fun, without compromising on performance, determinism, or distributed safety. At its core, SlugKit is a C++ service (built with userver) that turns structured patterns into unique, human-readable IDs — slugs like brave-tiger-7f3a, CozyFox42, or even roaring.stegosaurus.XVII. Each pattern defines a bounded domain of possible values, and SlugKit guarantees uniqueness through a deterministic series engine backed by atomic counters and a bit of number theory. It scales across nodes, supports streaming, and generates 50–300K slugs/sec per core, depending on pattern complexity. It’s fast. It’s safe. And it’s weirdly delightful to work with. SlugKit supports: • Custom pattern language ({adjective}-{noun}-{number:4x}, etc.) • Deterministic and random generation via mint, slice, and forge • Tag filters (+tech-nsfw,<10 chars, etc.) • Built-in slugginess prevention (avoids boring or repetitive outputs) • A dead-simple HTTP/JSON API and CLI • Also it provides and MCP server for AI agents to use collision-free sequences. It powers user handles, SKU naming, distributed ID coordination, and A/B test bucketing — anywhere a memorable ID matters. There’s a SaaS version with projects and org scopes, and an on-prem version you can Dockerize and run offline (well, almost — license checks still happen). And yes, there’s math behind the madness: slug domains are sized using prime-reduced LCMs for maximum cycle length and minimal collisions. This project wasn’t born in a boardroom. It was forged in late nights, debug logs, and too many coffee-fueled naming debates. Now it’s out there. Naming things. Better.