LM-Kit One

LM-Kit One is a private AI application server that allows users to run models, agents, retrieval-augmented generation (RAG), search, and document intelligence within their own infrastructure. It supports compatibility with OpenAI, Anthropic, Ollama, and MCP clients, offers comprehensive AI backend capabilities in a single installable package, and provides secure, scalable, and controlled operations with features like local storage, air-gapped modes, multi-tenant management, and fine-tuning.