About Promptention

Founded in 2024, Promptention is an AI security company. We work where AI systems meet real attackers: the models teams download, the agents they give tools and data to, and every request that passes between them. What we learn there becomes defenses teams can run in production.

We are a small team working across AI research and cybersecurity. That work informs Guard for runtime defense, Red for adversarial testing, and Model Clearance for checking downloaded model files before they are loaded.

We publish our thinking, test ideas, and welcome discussions with people working on the same problems.

Join our talent pool

If you are researching adversarial AI or model file security, we would like to learn about your work. Explore our research interests.