About Blue Sail Technologies
Blue Sail Technologies is an AI product company founded by Anthony McAniff in New York. We build AI-native products and run every layer beneath them ourselves.
Anthony McAniff
Founder & Engineer
Anthony is a DevOps and infrastructure engineer who is all-in on AI. He holds a Master of Science in Computer Science from Stevens Institute of Technology and brings deep experience in cloud-native platforms, Kubernetes, and building production systems from the ground up.
At Blue Sail Technologies, Anthony designs, builds and runs every product and the platform beneath them: a 9-node ARM64 Kubernetes cluster with three control-plane nodes, five workers and an NVIDIA DGX Spark that serves the company's local AI model, Gemma 4. Storage runs on Rook-Ceph, a 3-node HashiCorp Vault cluster handles secrets, and Cloudflare sits at the edge.
His approach is security-first and ship-it-to-production. Every project goes from idea to a live deployment on that infrastructure, with image scanning and TLS on every public site — and this site names the models behind our main products.
When he's not attached to a keyboard and monitor, Anthony spends his time body surfing in the Atlantic Ocean and enjoying time with his large extended family.
How we build
AI at the core
Each product is designed around what the model does well — writing, translating, explaining — not a chat box added at the end.
Local-first AI
Our products' local AI runs on hardware we own and operate. When a product also uses a cloud model, we name it.
Built and run in-house
Our products run on infrastructure we build, monitor and patch ourselves.
What we build
- ask-amacNew York's first AI-powered tourist guideFeatured
- The Clam LifeAI trends, explained
- Ukrainian courseLearn Ukrainian with stories written for your levelNew
- Learn any languageAny language, the same wayComing soon
- Job RabbitAI coach for résumés and cover letters
- Clam BookA coding and writing assistant in a 3D coral reef
- QT VaultQuantum-safe secrets for KubernetesEarly release
- GoNADGoNAD is Not Another DashboardIn development
Our platform
See how the platform worksCompute
9 ARM64 nodes: 3 control-plane, 5 workers and an NVIDIA DGX Spark AI worker, with 84 cores and about 370 GiB of memory.
AI
Gemma 4 26B-A4B on the DGX Spark behind our in-house gateway, with Llama Guard 3 and embedding models alongside.
Storage
13 TiB of raw NVMe in a Rook-Ceph cluster, plus a 3-node Qdrant vector database.
Edge and security
Cloudflare in front, automatic Let's Encrypt certificates, HashiCorp Vault and Prometheus alerting.
Tell us what you're building.
A product idea, a private AI setup for your team, or feedback on something we've made. We typically respond within 24 hours.