manifesto

Where copy learns.

From marketing emails to product descriptions, from ads on your feed to billboards on the street, from political campaigns to product launches – everything runs on text.

Text is how ideas travel, how we sell, how we persuade, how we move the world.

Every word you read has been crafted by a human, for a human to make you think, feel, or act.

But it’s static. It’s generic. It’s one-size-fits-all.

We think that’s insane. And we’re here to change it.

We want text to adapt, to specialize, to learn, to understand you.

To evolve until every word is written not just for a crowd, but for a person.

That’s what inkRL is building: a system that transforms language from something designed by humans for humans into something optimized with them: a world where words don’t just speak, but learn.

Copy that adapts itself in real time. Feedback loops stitched into every sentence. Reinforcement learning built for language that listens.

The Vision

We imagine a world where every single word, every line of copy, every subject line, every CTA learns from how people react to it.

A world where language isn’t static, but alive. Where algorithms don’t just generate text, but understand and improve it, continuously, autonomously, and at scale.

Our goal is to build the algorithm that personalizes communication for every individual. An algorithm that learns what resonates with each reader, and optimizes towards one universal metric: engagement.

We’re starting with email: the most measurable, most iterative marketing channel. But we’re building toward something bigger: a world where every piece of text is optimized for its goal.

The Belief

We believe creativity can be explainable.

We believe preferences can be embedded, tracked, and learned.

We believe reinforcement learning, the same feedback loop that mastered games, can master persuasion.

TikTok and Meta already do this for feeds and reccomandation. We’re doing it for language.

The Mission

inkRL is building a proprietary reinforcement learning system that autonomously improves marketing copy to maximize user engagement.

Every campaign becomes a feedback loop. Every click teaches the model something new. Every word is an experiment.

We’re creating a personalized optimization layer for the internet where text adapts to humans, not the other way around.

inkrl

Reinforcement learning for language that learns you back.