Frequently Asked Questions

Shakudo is the sovereign AI orchestration platform for enterprises that cannot ship their data outside their own infrastructure. It sits underneath the AI stack and keeps every layer of it, models, data, agents, tools, running inside the customer's own perimeter. The whole platform is built on one conviction: that the companies doing the most valuable AI work will be the ones where the data never leaves.

Sovereign AI means the data stays on your own infrastructure, under your control. The model, the training, the inference, all of it runs inside your perimeter. It is not a product. It is the operating principle underneath the product.

Because the moment you send sensitive data to a third party cloud, you have handed over control of your most valuable asset. Some industries cannot do that. Not because of fear, but because the work itself is regulated. Sovereign AI is the answer to that constraint.

Kaji runs inside your own infrastructure, not in a vendor's cloud. Every action it takes, every memory it forms, every skill it acquires, is stored on your systems. It is not a chatbot pointed at an API. It is an operating layer for agentic workflows that understands your environment because it lives in it.

That is the question we get most often, and the honest answer is: almost anything that can be expressed as a workflow. Writing code and pushing it to a repo. Checking a customer's status in your CRM and updating it. Pulling data from a database, running an analysis, and returning a summary. The boundary is not the model. It is the skills you give it and the access you grant.

Kaji connects through skills, which are instruction sets paired with the tools it needs. A skill can be a shell command, an API call, a browser session, a script. You define what Kaji can do, and it learns to use those tools to complete tasks. There is no lock in to a vendor API. If you can script it, Kaji can do it.

Yes. Kaji is model agnostic by design. It works with open source models, with models you have fine tuned, and with models running on your own GPU infrastructure. That is a core part of the sovereignty story. You are not locked to a single provider's model, and your data never has to leave your environment to reach one.

You run it on your own infrastructure. That could be a GPU server in your office, a Kubernetes cluster you already operate, or a dedicated air gapped environment. The platform is containerized and designed to run in whatever environment you already trust. There is no cloud dependency. If you can run a container, you can run Shakudo.

It does not. That is the point. The data never leaves your perimeter, so the compliance burden is fundamentally different. You are not asking a cloud vendor to handle your PHI or your financial data. You are running the whole pipeline on systems you already control and already audit. For regulated industries, that changes what is possible.

The short answer is: you define the boundary, and the platform enforces it. Kaji operates within the scope of the skills and tools you give it. It does not have open access to your environment by default. And because it runs on your infrastructure, your security team can see, audit, and control every action it takes. That level of oversight is something you do not get from a SaaS product.

Kaji maintains a persistent memory of context, decisions, and relationships it encounters. It does not start from zero every time you open a session. It remembers the project, the people involved, the decisions already made. That memory is stored on your infrastructure, not in a vendor database. It is what makes Kaji feel less like a tool and more like a colleague that has been in the room.

Those are developer frameworks. They help you build agentic workflows in code. Shakudo is an operating platform. It gives you the infrastructure, the memory, the security layer, the deployment story, and the human interface all in one system. You can use a framework underneath if you want to, but the platform is the thing that makes the whole stack run inside your perimeter and stay manageable at enterprise scale.

No.

Depends on your environment, but a basic deployment is typically days, not months. The platform is designed to be deployed on existing Kubernetes or container infrastructure. The complexity is in the skills and integrations you build, not in getting the core platform running.

Pricing is usage based and depends on the infrastructure you deploy on and the scale of the agentic workloads you run. There is no per-seat licensing model, because the product is not a per-person SaaS tool. Talk to the team for a quote that reflects your specific use case.

Anywhere the data cannot leave. That is healthcare, finance, energy, government, defense, and increasingly, any enterprise with serious IP concerns. If you are building AI workflows on data that a cloud vendor would refuse to touch, you are in the right category.

Shakudo does not replace your models. It gives you the operating layer to run them, secure them, and give your people a way to use them without writing code. If you have already invested in fine tuning models or building GPU infrastructure, Shakudo is the platform that makes that investment usable at scale across your organization.

Yes. That is one of the core design requirements. The platform runs fully offline. No external API calls, no cloud dependencies, no phone home. If your environment is air gapped, Shakudo runs there.

Enterprises that have real data, real constraints, and a reason to keep both inside their own walls. That is not every company. But for the ones that are, the question is not whether to build sovereign AI. It is whether to build it with a platform purpose built for the job or to try to assemble it from parts.

The direction is clear: agents that can plan, execute, and verify multi step workflows across an organization's systems without human intervention for every action. The constraint is not the model. It is trust. You need to know what the agent is doing, where the data is, and that nothing is leaving your perimeter. That is the problem Shakudo was built to solve, and it is the one that will define which AI systems enterprises actually deploy at scale.