Private AI, deployed on your vessel
Every engagement starts with a conversation about what you actually want to accomplish, and whether we are the right team to build it. Below are the shapes the work usually takes. Your situation may be one of these, a combination, or something we scope together.
A note on scope: we build the AI and data software layer, and we install the compute that runs it. We are not a classification society, a shipyard, or a bridge-equipment OEM. Safety-of-life and class-approved systems are work we coordinate with the parties qualified for them, not work we sell directly.
A private ChatGPT, running on your vessel
Private LLM & Knowledge
Open-weights language models deployed on local hardware, with retrieval over the documents, manuals, and data that matter to your operation. Conversations, answers, and assistance that stay aboard. No cloud account, no vendor data lake, no audit trail outside your control.
- Local model serving (Llama, Mistral, Qwen, Kimi families)
- Retrieval over vessel-specific corpora
- Conversational interfaces for owner, guests, and crew
- Kill switch and data-wipe controls
Call your onboard AI from any phone
Voice AI
A voice interface that answers when the satellite link is degraded or fully down. Reach the onboard AI from any handset, intercom, or line, and hear useful answers grounded in the vessel's own knowledge base. The signature of how we work.
- Phone and intercom integrations with low-latency speech
- Grounded answers from the vessel's own documents
- Works fully offline once deployed
- Optional call logging with owner-controlled retention
Set up the server room to run it
On-Vessel Compute
Specification, installation, and ongoing operation of the compute that serves your models. We work with the server space you have or help you scope new, and we are deliberate about not pretending to be a shipyard integrator. Class-approved hardware and structural work sits with the parties who are qualified for it.
- GPU server spec and setup (NVIDIA H100, H200, L40S, B200)
- Model serving, monitoring, and over-the-air updates
- Privacy controls: at-rest encryption, kill switch, offline mode
- Works with existing IT infrastructure or helps you refresh it
What the consultation scopes, we build
Custom Applications
Most of the interesting work is bespoke. Guest concierge, crew knowledge base, operations co-pilot for ETOs, automations that tie the AI into CRMs, telephony, or internal tools. Scoped during the consultation, built to fit, and never outside a sober read of what we can actually deliver.
- Agentic workflows against existing business systems
- Integrations with CRMs, calendars, telephony, and APIs
- Owner / crew / guest experiences tuned to a specific vessel
- Clear scope of work, no drift into safety-critical territory
Figure out what is actually worth building
Consulting & Strategy
Assessment, architecture design, vendor evaluation, and honest roadmaps from engineers who have shipped local AI systems. Our first conversation is always whether the thing you are thinking about is worth doing at all. Sometimes the answer is no, and we will say so.
- AI readiness assessment for a vessel or fleet
- Architecture and capacity planning for local compute
- Vendor and model evaluation
- Honest roadmaps with scope boundaries written in
How we work
A deliberate, sequential process. No shortcuts, no shore-side assumptions.
01
Conversation
We start with a call. We listen for what you actually want to accomplish, what data lives where, and which constraints matter. Half of these conversations end with us telling you a smaller scope will serve you better.
02
Design
We write a scoped proposal. Hardware we would install or spec with your existing IT, models we would run, interfaces we would build, clear boundaries on what we do and do not cover. No surprises.
03
Build and Deploy
We build the software and set up the compute. Where vessel work requires a shipyard, class society, or certified marine electrician, we coordinate. We do not pretend to be those parties.
04
Hand-Off and Support
We document the system so your team can live with it. We remain available for ongoing updates and operational support on terms that match your uptime needs, not ours.
Own your AI. Own your data. Keep working when the link drops.
Whether you're a single superyacht or a fleet of cruise ships, we design and deploy sovereign AI that stays yours: on the vessel, off the cloud, under your control.