About this skill

Problem Solved

In agent workflows, tasks often require gathering public information from multiple search engines, but single-page scraping is limited and paid APIs introduce keys and cost. multi-search-plus provides a lightweight, API-key-free search approach that targets public URLs and focuses on Chinese web sources plus common international engines.

How It Works

It supports 17 engines, including baidu, bing_cn, sogou_wechat, google, duckduckgo, brave, and wolframalpha.
It offers two paths: an LLM can call web_fetch for a single-engine URL during conversation, or run scripts/multi_search.py for batch parallel scheduling with a default 3-worker pool and 1.5s delay.
Output can be --json, --md, or --tsv, with URL normalization and utm_* parameter deduplication to support downstream analysis.
Search operators such as site:, filetype:, exact matching, exclusions, and OR queries are supported, and Google results can be time-filtered with --time week.

Boundaries

It is better suited for one-off research, cross-engine comparison, Chinese-language source discovery, and structured output generation than for continuous monitoring or high-volume harvesting. Page redesigns, rate limits, and ToS constraints can affect stability, and frequent requests should respect rate limits. For privacy-sensitive queries, prefer duckduckgo, startpage, and brave rather than mixing in Google by default.

Screenshots

Use Cases

When researching an industry white paper, query Baidu, Bing China, Sogou, and Toutiao, then output a Markdown summary with sources.
When looking for a brand's WeChat articles, use sogou_wechat to search public WeChat links and deduplicate URLs.
When compiling sourcing research, run batch_search.py across multiple keywords and sites, then export JSON for analysis.
When reviewing an open-source framework, search Google HK, Brave, and DuckDuckGo with site: or filetype: filters.

Best For

Product analysts doing Chinese market research who need competitor materials from multiple domestic engines and sourced summaries.
Content operators researching brand topics who need WeChat articles, Toutiao discussions, and Zhihu content for topic selection.
Data engineers who need multi-engine search results converted to JSON or TSV for deduplication and field analysis.
Technical researchers who need international engines, GitHub repositories, PDF files, and tech blogs for side-by-side source comparison.