Integrate Baidu AI Search for Smarter Answers

This MCP server exposes Baidu's AI Search component, which combines Baidu's search capabilities with large language model technology to produce intelligent responses backed by real-time web information references. It is built for a range of industry application scenarios and offers several standardized, configurable capabilities: custom persona settings, model selection, query rewriting (including time-sensitive and multi-turn rewriting to improve search results), search scope configuration (choice of modalities, site ranges, and publication dates), and a customizable number of reference links returned with each response.

Use it when an application needs LLM-generated answers grounded in current web search results, with control over which sites, time windows, or content modalities the search draws from, and how many reference links come back with each answer. It is not a general-purpose Baidu Cloud connector - it is scoped to the AI Search capability specifically, configured through Baidu's AppBuilder console.

Configurable options include a custom assistant persona, choice of underlying model, query rewriting tuned for time-sensitive or multi-turn conversations, restricting search scope by modality, site range, or publication date, and setting how many reference links are surfaced per response.

Get an AppBuilder API Key from the Baidu AppBuilder console, then format the authorization token as Bearer+<AppBuilder API Key> (keeping the "+" between the two parts). Configure the MCP client with:

{
  "mcpServers": {
    "baidu_ai_search": {
      "url": "http://appbuilder.baidu.com/v2/ai_search/mcp/sse?api_key=Bearer+bce-v3/ALTAK..."
    }
  }
}

Further documentation is available in Baidu's official AI Search docs (cloud.baidu.com).

Developers building on Baidu's AppBuilder platform who want search-grounded LLM answers with fine-grained control over search scope, persona, and reference count, for Chinese-market or Baidu-ecosystem applications.