ai infrastructure · keisei-sdk

KeiSeiSDK — standalone agent runtime.

A vertical-stack agent runtime and CLI for self-hosted LLM deployments. Runs against locally-served models or remote endpoints, with the same agent semantics across both.

What it is

KeiSeiSDK is what the lab uses when an agent needs to run end-to-end on its own — outside a hosted IDE, against a local or remote LLM. CLI-first, scriptable, with the same agent definitions and skill surface as the rest of the KeiSeiKit family.

Why standalone

Embedding an agent inside another tool is fine for development. Running an agent as its own first-class process is what makes it deployable on a server, a workstation, or an embedded box.