About Katib

Katib is the system of record for editorial thinking — a memory and workflow layer that sits above the foundation models, not another AI writer. It’s built for Arabic creators, founders, and small teams who want content that sounds like them, not like the model everyone else is using too.

What we believe

The model is a commodity. Anyone can pipe a prompt into the same endpoint and get the same fluent, forgettable draft. What can never be commoditised is your editorial judgement — the angle you chose, the line you cut, the point of view you’ve spent years sharpening.

So Katib doesn’t try to write better than you. It captures how you decide. The defensible asset isn’t the generation; it’s the owned, evolving voice and the compounding record of every choice that voice has made.

How we work

The product is a loop. You capture an ideabefore it escapes. In the Editorial Lab you generate several angles and read them side by side — discovery, operational, contrarian — then keep the strongest and edit it in place, instead of one-shotting a single prompt. Your voice is a structured asset, not a prompt: a declared layer you set, and a learned layer that sharpens every time you pick one angle over another.

Every decision lands in the Archive — an immutable record of what you kept, what you cut, and the version of your voice that shaped it. Re-open any decision and the voice that made it is pinned in place; a later change never rewrites the record. The model proposes; you decide; we remember.

Where we’re going

Phase 2 brings multi-user collaboration on a shared voice and decision record, and a richer Archive you can search and learn from. The model behind the gateway will keep changing; your voice and your record won’t. If you publish in Arabic and want to shape what we build next, get in touch.