MaxKB adopts an open model docking architecture and supports three main types of model access:
- Public Cloud ModelIncluding OpenAI GPT series, Anthropic Claude, Smart Spectrum AI and other domestic and international mainstream APIs.
- open source local model: Support for self-hosted models such as Llama 3, ChatGLM3, Qwen, etc. accessed via Ollama or vLLM
- Enterprise-specific models: Compatible with any localized deployment model that conforms to the OpenAI API specification
Specific steps for configuring a local private model:
- Select the "Customize Model" option in the Model Management module.
- Fill in the model API address (e.g. http://localhost:11434/api/generate)
- Set parameters such as model name and context length
- Save configuration after performing connectivity tests
The system also supports model hot-switching and AB testing, which can automatically assign the most suitable model according to different business scenarios. For example, customer service scenarios can use the lower-cost 7B small model, while technical document parsing can be switched to a large model with 70B parameters.
This answer comes from the articleMaxKB: Out-of-the-box AI Knowledge Base Q&A System for Smart Customer Service and In-house Knowledge BaseThe
May not be reproduced without permission:AI productivity tools " What types of large language models does MaxKB support docking? How to configure a local private model?
