Bocha Search MCP Server

Bocha Search MCP is a purpose‑built search engine for AI assistants, offering developers instant access to nearly ten billion web pages and a wide array of curated content such as weather, news, encyclopedias, medical information, train schedules, and more. Unlike traditional keyword‑centric search APIs that prioritize click‑through revenue, Bocha delivers clean, high‑quality answers tailored for conversational AI. Its semantic ranking engine, powered by Transformer models, evaluates relevance between a user’s query and candidate pages, ensuring that the most context‑appropriate results surface at the top of each response.

At its core, the server exposes two complementary tools. Bocha Web Search performs a broad web crawl and returns structured metadata—titles, URLs, summaries, publication dates, site icons, and image links—for up to fifty results per query. Bocha AI Search extends this capability by interpreting the intent behind a natural‑language request and augmenting the output with modal cards—rich, domain‑specific data structures such as weather forecasts, calendar events, encyclopedic entries, stock prices, and medical facts. These modal cards provide instant, actionable information that a large language model can ingest directly, dramatically reducing hallucination and improving answer freshness.

For developers building AI workflows, Bocha integrates seamlessly with popular MCP‑compatible clients like Claude Desktop, Cursor, and the Ark platform. By configuring a single environment variable with an API key from the Bocha AI Open Platform, developers can invoke search or AI‑enhanced search directly within their assistant’s prompt logic. The API’s flexible freshness filters (e.g., , ) allow fine‑grained control over result timeliness, making it ideal for time‑sensitive applications such as travel booking assistants or real‑time financial advisors.

Real‑world use cases span from powering the search backbone of major enterprises—DeepSeek, Alibaba, Tencent, and ByteDance already rely on Bocha—to enabling niche bots that answer medical queries or provide up‑to‑date train schedules. Because Bocha’s semantic ranking aligns closely with Transformer‑based models, the returned results are naturally favored by large language models, leading to more accurate and contextually relevant conversations. Its competitive pricing (under $0.02 per request) and domestic data coverage also make it a compelling alternative to costly overseas APIs like Bing Search.

In summary, Bocha Search MCP delivers high‑quality, semantically ranked web and domain data directly to AI assistants. Its dual‑tool architecture—plain web search plus AI‑enhanced modal cards—provides developers with a versatile, low‑latency solution for integrating real‑world knowledge into conversational experiences.