MaxKB: Enhancing AI Applications with Real-time Search

MaxKB: Enhancing AI Applications with Real-time Search

What is MaxKB?

MaxKB (Max Knowledge Base) is a ready-to-use, flexible RAG chatbot system built on Large Language Models (LLM) and Retrieval-Augmented Generation (RAG) technology. As a mature knowledge base Q&A system, MaxKB is widely used in scenarios such as intelligent customer service, corporate knowledge management, academic research, and education.

How Does RAG Technology Improve AI Application Quality?

Traditional large language models often produce "hallucinations" when answering questions - generating content that seems reasonable but is actually inaccurate - due to knowledge limitations or outdated training data. Retrieval-Augmented Generation (RAG) technology addresses this issue by first retrieving relevant information from reliable knowledge bases before generating answers, then using these retrieval results to guide the model toward more accurate responses, effectively reducing hallucinations.

By implementing RAG technology, MaxKB enables AI applications to:

  1. Access real-time information: Overcoming the limitations of model knowledge cutoff dates
  2. Provide traceable answers: Every response has clear knowledge sources
  3. Generate more accurate content: Reducing hallucinations and misinformation

MaxKB's Comprehensive Real-time Search Enhancement Capabilities

Web Search Integration: Beyond Static Knowledge Bases

The latest version of MaxKB offers breakthrough web search capabilities, allowing AI applications to access the most current information from the internet:

  1. Google Search Integration: Through the built-in Google Web Search function, MaxKB can directly retrieve public information from the internet, accessing the latest web content and data.
  2. LangSearch API: A natural language web search API that enables the system to understand user queries in natural language and retrieve relevant information from the internet.
  3. Sogou AI Integration: Retrieves information and webpage URLs from the Sogou search engine, expanding information source channels.
  4. PostgreSQL and MySQL Queries: Supports direct database queries, enabling real-time connections with enterprise data systems.

These web search functions can be flexibly orchestrated through MaxKB's powerful workflow engine, seamlessly integrating with knowledge base retrieval, LLM Q&A, and other nodes. This gives AI applications the ability to obtain and process internet information in real-time, significantly improving the timeliness and accuracy of responses.

Multiple Retrieval Mode Support

In addition to web search, MaxKB supports three powerful knowledge base retrieval methods to ensure AI can find the most relevant information:

  1. Vector Retrieval: Calculates the similarity between user questions and text segments using vector models to identify semantically similar content.
  2. Full-text Retrieval: Uses keyword searches to return text segments containing the most keywords.
  3. Hybrid Retrieval: Performs both full-text and vector retrieval, then reranks results to select the best matches from both query types.

Knowledge Base Diversity

MaxKB supports two types of knowledge bases:

  1. General Knowledge Base: Supports offline documents, including various text files, tables, and Q&A pairs.
  2. Website Knowledge Base: For online static text data, automatically synchronizes text data from root URLs and sub-URLs after inputting the root web address.

Supported document formats include Markdown, TXT, PDF, DOCX, HTML, XLS, XLSX, CSV, ZIP, and more, meeting various enterprise data integration needs.

Workflow-based Intelligent Orchestration

MaxKB features a powerful workflow engine that allows users to orchestrate complex AI processing workflows through a visual interface:

  1. Web Search Nodes: Access real-time internet information through built-in Google Search, LangSearch API, and other functions.
  2. Knowledge Retrieval Nodes: Connect to knowledge bases to retrieve relevant segments for questions.
  3. Multi-path Recall: Use reranking models to perform secondary recall on search results from multiple knowledge bases.
  4. Conditional Branches: Execute different nodes based on varying conditions.
  5. Question Optimization: Optimize user questions through AI to improve retrieval precision.

Fine-grained Search Control

MaxKB provides rich parameter settings for precise control over the search process:

  1. Similarity Threshold: Controls the relevance requirements between retrieval results and questions.
  2. Top-N References: Returns the top N segments or search results that meet the criteria.
  3. Maximum Reference Characters: Controls the length of referenced content.
  4. Multi-source Information Integration Strategy: Allows configuration of how web search and knowledge base retrieval results are integrated.

Application Scenarios for Real-time Search Enhancement

Real-time Information Services

Leveraging MaxKB's web search capabilities, organizations can build real-time information services providing the latest news, weather, stock prices, product pricing, and more, no longer limited by the timeliness constraints of static knowledge bases.

Research and Innovation Support

Researchers can use MaxKB's web search capabilities to access the latest research papers, technological developments, and market trends, accelerating the innovation process. Simultaneously, they can retrieve historical research materials from internal knowledge bases, effectively integrating internal and external information.

Advanced Customer Service

Businesses can combine product knowledge bases with real-time web searches to not only answer product-related questions but also provide up-to-date comparisons with market competitors, user reviews, and usage tips, significantly enhancing the customer service experience.

Intelligent Decision Support

Managers can use MaxKB to simultaneously search enterprise internal databases and internet information, obtaining comprehensive decision-making evidence including market trends, competitor movements, and internal business data to support strategic decisions.

Education and Learning Platforms

Educational institutions can combine curriculum content with the latest internet resources to provide students with a comprehensive learning experience that includes both structured course knowledge and real-time industry developments and application cases.

MaxKB's Technical Advantages

  1. Real-time Web Connectivity: Achieves real-time connections to the internet through built-in search API interfaces, obtaining the latest information.
  2. Ready-to-Use: Supports direct document uploads/automatic web crawling, with automatic text segmentation, vectorization, and RAG processing.
  3. Flexible Orchestration: Equipped with a powerful workflow engine and function library, enabling flexible combinations of various search nodes to meet complex business scenario requirements.
  4. Seamless Integration: Enables zero-code rapid integration with third-party business systems, quickly empowering existing systems with intelligent Q&A capabilities.
  5. Model-Agnostic: Supports various large models, including private models (such as DeepSeek, Llama, Qwen, etc.) and public models (like OpenAI, Claude, Gemini, etc.).

Conclusion

Through the dual enhancement of web search and knowledge base retrieval, MaxKB breaks through the limitations of closed knowledge systems. Whether requiring the latest internet information or proprietary internal knowledge, MaxKB can provide precise, reliable, and real-time intelligent services through flexible orchestration and powerful search capabilities. This comprehensive real-time search enhancement capability provides a solid technical foundation for enterprise intelligent transformation and AI application deployment.

Back to blog