Foundry captures what you say and read, builds it into a living memory, and lets you — and your AI — act on it. Bring your own AI, or use ours.
Every plan runs on your own private cell — isolated storage, encrypted before it's ever written, your data never trains anyone's model.
You get a generous monthly allowance on capable models. You only pay more if you run far past it, or reach for the most powerful frontier models.
Each plan bundles a monthly pool of AI on efficient, capable models — enough for everyday questions, summaries, and playbook runs.
Run the included models within your allowance, meter stronger models by the token, or bring your own key for Claude or GPT and pay us nothing for that inference.
Past your allowance, extra usage is billed per million tokens — at the model's rate plus a small margin. Most people never reach it.