1Panel-dev/MaxKB

What it solves

MaxKB provides a unified platform for businesses to build enterprise-grade AI agents. It addresses the challenge of reducing hallucinations in large language models by integrating Retrieval-Augmented Generation (RAG) and allows for the creation of complex business workflows without requiring extensive coding.

How it works

MaxKB uses a RAG pipeline that allows users to upload documents or crawl online content, which is then automatically split and vectorized. It leverages a workflow engine and MCP tool-use capabilities to orchestrate AI processes. The system is model-agnostic, supporting both public models (e.g., OpenAI, Claude) and private models (e.g., DeepSeek, Llama), and is built on a technical stack including LangChain, PostgreSQL with pgvector, and Django.

Who it’s for

It is designed for organizations needing intelligent customer service, corporate internal knowledge bases, academic research, and educational tools.

Highlights

• RAG Pipeline: Automatic text splitting and vectorization of uploaded or crawled documents.
• Agentic Workflow: A powerful engine for orchestrating AI processes using a function library and MCP tool-use.
• Zero-Coding Integration: Rapid integration into third-party business systems for intelligent Q&A.
• Model-Agnostic: Compatible with a wide range of public and private LLMs.
• Multimodal Support: Native support for text, image, audio, and video input and output.