Higress AI-Search Server

Example Use Cases

• 1

  Augment LLM outputs with up-to-date web search results (Google, Bing, Quark) to answer current events or general knowledge queries.

• 2

  Pull academic paper metadata and summaries from Arxiv to support research or literature-aware responses.

• 3

  Surface internal company knowledge (employee handbooks, policies, process documents) alongside web search results for enterprise Q&A.

Description

This repository provides a Model Context Protocol (MCP) server that integrates Higress ai-search and ai-proxy plugins to fetch live search results (Google, Bing, Quark, Arxiv) and inject them into model responses. It can be configured via environment variables and run locally or via uv/uvx for easy deployment. The server is intended to enrich conversational agents with up-to-date external and internal knowledge (employee handbooks, policies, documents) to improve answer relevance and accuracy.