Bocha Wanxiang is a search and semantic-ranking platform built for AI applications. Its Web Search API and Semantic Reranker API add external knowledge retrieval to agents, chatbots, AI search products, and RAG systems. Natural-language queries can retrieve results from web pages and content categories such as news, images, video, encyclopedic material, travel inventory, and academic sources, making the service primarily an integration layer for developers rather than a general consumer search page.
Teams can connect through the official API, model tool calls, or plugins on supported agent-development platforms, with options for custom filters and streaming or non-streaming responses. The reranking service works on an existing set of candidate documents and reorders them according to semantic relevance to the query. Calling the official endpoints requires an API key sent with Bearer authentication, while result verification, presentation, and downstream application behavior remain the integrating team’s responsibility.