We build AI products that run on your phone — for travel, money, and everyday life.
Two products. One AI engine. Runs on your device.
Your AI assistant, on your phone
A private AI assistant that thinks on your phone. It uses Apple Intelligence and open-weight models so your data stays with you. When a task needs more horsepower, it falls back to cloud models like Claude Sonnet, Llama 4, or DeepSeek.
Smart travel for the Americas
Lanzo finds hidden flight routes, optimizes your loyalty points across borders, and tells you which credit card to use — before you book.
Next up: Ombra — an AI agent that handles video calls and messages on your behalf.
The engineering behind the products.
AI that runs privately on your phone without sending data to the cloud. We use Apple Intelligence and open-weight models to keep things fast, personal, and offline-capable.
Our backend runs at the edge — close to users, globally distributed, and fast. When heavier compute is needed, we scale up without making you wait.
Built infrastructure for AI safety labs evaluating frontier models. Contributed to METR's open-source Inspect framework used to measure what advanced AI systems can actually do.
Before SpookyFox, we built AI infrastructure and shipped products at places like these.
Built the compute platform behind Inspect, an open-source framework used by AI safety labs to evaluate frontier model capabilities.
Co-founded a 700-member non-profit community funding open-source AI and developer tools through grants and collaboration.
Shipped AI-integrated products with production-grade identity security at Okta and research infrastructure at UC Berkeley.