1. Project Overview

A knowledge base question-answering system powered by LLMs

MaxKB is a knowledge base question-answering system powered by LLMs. MaxKB = Max Knowledge Base, aiming to become the most powerful brain for enterprises.

2. Open Source License

Licensed under GPL-3.0

3. Screenshots

UI Showcase

4. Feature Overview

• Ready to use out of the box: Supports direct document uploads and automatic crawling of online documents, with automatic text splitting, vectorization, RAG (Retrieval-Augmented Generation), and an excellent intelligent question-answering experience;

• Seamless integration: Supports rapid, zero-code integration into third-party business systems;

• Multi-model support: Supports integration with mainstream large language models, including locally hosted private models via Ollama (such as Meta Llama 3, qwen, etc.), Tongyi Qianwen, OpenAI, Azure OpenAI, Kimi, Zhipu AI, iFlytek Spark, Baidu Qianfan, and more.

Quick Start

docker run -d --name=maxkb -p 8080:8080 -v ~/.maxkb:/var/lib/postgresql/data 1panel/maxkb

# Username: admin
# Password: MaxKB@123..

You can also quickly deploy MaxKB + Ollama + Llama 2 through the 1Panel App Store. Within 30 minutes, you can launch a knowledge base question-answering system powered by local large language models and integrate it into third-party business systems.

For offline network environments, we recommend using the offline installation package for installation and deployment.

You can also try it online: DataEase Assistant, an intelligent question-answering system built with MaxKB and integrated into the DataEase product and online documentation.

5. Tech Stack

Tech Stack

• Frontend: Vue.js

• Backend: Python / Django

• LangChain：LangChain

• Vector database: PostgreSQL / pgvector

• Large language models: Azure OpenAI, OpenAI, Baidu Qianfan, Ollama, Tongyi Qianwen, Kimi, Zhipu AI, iFlytek Spark

1. Installation and Deployment

1.1 System Requirements

Before installation, ensure that your system meets the following requirements:

• Operating System: Ubuntu 22.04 / CentOS 7 64-bit;

• CPU/Memory: 2C/4GB or higher recommended;

• Disk Space: 100GB;

• Browser Requirements: Please use modern browsers such as Chrome, FireFox, Safari, Edge, etc.;

• Internet access is required.

1.2 Quick Online Deployment

1. MaxKB supports one-click startup. Simply execute the following command:

docker run -d --name=maxkb -p 8080:8080 -v ~/.maxkb:/var/lib/postgresql/data 1panel/maxkb

2. After successful installation, access MaxKB through your browser:

http://Target Server IP Address:Target Port

Default Login Credentials
Username: admin
Default Password:MaxKB@123..

If you are using a cloud server, please open the target port in the security group.

1.3 Offline Deployment (Recommended for Production Environments)

Note: The offline package only supports x86 servers.

1. Download the MaxKB offline package from the community website: Community Edition Offline Package

2. Upload the package to the server, extract it, and execute the following command:

# Replace maxkb-v1.2.0-offline.tar.gz with the name of the downloaded package
tar -zxvf maxkb-v1.2.0-offline.tar.gz

3. Install MaxKB by executing the following commands:

# Navigate to the extracted installation package directory
cd maxkb-v1.2.0-offline

# Execute the installation command
bash install.sh

4. After successful installation, access MaxKB through your browser:

http://Target Server IP Address:8080

Default Login Credentials
Username: admin
Default Password:MaxKB@123..

More content, please refer to README.md document