Minima — локальный RAG-поиск по документам

★ 1K

Minima is an open-source RAG in containers for on-premises deployment: it indexes your documents and lets you answer questions over them — fully locally or paired with ChatGPT and MCP. Reach for it when you want private question-answering over your documents with flexible control of where the LLM runs: fully isolated on your own hardware (via Ollama), with your own LLM over an OpenAI-compatible API (vLLM/TGI, etc.), or with a local indexer while ChatGPT remains the primary model through a custom GPT. It deploys in containers, and your data and index stay with you. Its focus is a flexible self-hosted RAG over your documents with several privacy modes, not a ready enterprise search platform with connectors (Onyx for that) or a vector database: the value is control over data and the choice of a local or external model. It suits people who want private document Q&A with the option to plug in ChatGPT/MCP as desired.