MaxKB

MaxKB is an open-source, self-hosted web app and server that provides a comprehensive platform for building enterprise-grade AI agents. It combines Retrieval-Augmented Generation (RAG) pipelines with workflow automation to manage corporate knowledge bases and automate customer service interactions. By integrating these components, the software allows organizations to build specialized assistants that rely on verified internal data rather than general training sets.

The software is deployed as a server-side web application, typically run via a Docker container. It allows users to create intelligent Q&A systems by connecting large language models to specific datasets, making it suitable for academic research, education, and internal corporate use. The system provides a web interface for managing knowledge sources, configuring agent behaviors, and deploying the resulting bots into production environments.

Key features

• RAG pipelines with document uploading and automatic web crawling
• Automatic text splitting and vectorization to reduce model hallucinations
• Workflow engine for orchestrating AI processes in complex business scenarios
• Support for MCP tool-use and a dedicated function library
• Compatibility with public models including OpenAI, Claude, and Gemini
• Support for private models such as DeepSeek, Llama, and Qwen
• Native multi-modal support for text, image, audio, and video input and output
• Zero-coding integration for adding Q&A capabilities to third-party systems

MaxKB is built using a Vue.js frontend and a Python Django backend. It utilizes the LangChain framework for LLM orchestration and relies on PostgreSQL with the pgvector extension for storing and querying vector embeddings. This architecture enables the platform to handle high-dimensional data efficiently while maintaining the flexibility to switch between different model providers. The platform is designed for organizations that need to maintain strict control over their data and infrastructure while leveraging both proprietary and local AI models.

This tool serves as a knowledge management layer that bridges the gap between raw enterprise data and generative AI interfaces, providing a structured way to deploy agentic AI in professional settings.