# Aas-ee/open-webSearch [Health: Active]

**Category:** 🔎 Search & Data Extraction  
**Repository:** https://github.com/Aas-ee/open-webSearch  
**GitHub Stars:** 1915  
**npm Downloads (last month):** 41387  
**Views:** 18  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/aas-ee-open-websearch

## Description
Web search using free multi-engine search (NO API KEYS REQUIRED) — Supports Bing, Baidu, DuckDuckGo, Brave, Exa, and CSDN.

## Tools
Capabilities this server exposes over MCP:

- **search** — Search the web using multiple engines (e.g., Baidu, Bing, DuckDuckGo, CSDN, Exa, Brave, Juejin(掘金), Startpage, Sogou(搜狗), Hacker News) with no API key required.
- **fetchLinuxDoArticle** — Fetch full article content from a linux.do post URL
- **fetchCsdnArticle** — Fetch full article content from a csdn post URL
- **fetchGithubReadme** — Fetch README content from a GitHub repository URL
- **fetchWebContent** — Fetch content from a public HTTP(S) URL. renderMode defaults to auto: request uses HTTP only, auto uses request with browser fallback, and browser renders directly with Playwright and fails clearly when Playwright is unavailable.
- **fetchJuejinArticle** — Fetch full article content from a Juejin(掘金) post URL

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "open-websearch": {
    "command": "npx",
    "args": ["-y","open-websearch@latest"]
  }
}
```

## Documentation

## What Aas-ee/open-webSearch does

Aas-ee/open-webSearch exposes web search and content retrieval through MCP, with no search API keys or authentication required. It can query multiple engines, including Bing, Baidu, DuckDuckGo, Brave, Exa, CSDN, Juejin, Startpage, Sogou, Hacker News, and Keeenable. The available engine set can be restricted through configuration, and a default engine can be selected.

The server also extracts content from selected sources. Dedicated operations handle Linux.do posts, CSDN articles, Juejin posts, and GitHub repository README files. A generic fetch operation accepts public HTTP(S) URLs and can return web or Markdown content.

## How it works

Run the package as an MCP process with `npx open-websearch@latest`, or configure it as a longer-lived local service using `open-websearch serve`. The package also provides a CLI for one-shot searches and a local HTTP daemon with health, search, and fetch endpoints.

Search results are returned in structured form with titles, URLs, and descriptions. For content research, a practical workflow is to search first and then fetch a specific result page. Generic fetching supports three rendering choices: HTTP requests only, automatic request-with-browser fallback, or direct Playwright rendering. Browser rendering requires Playwright to be available.

## Setup and configuration

No environment variables are required for basic operation. Optional settings control server mode, port, CORS, proxy use, search-engine selection, and fetch behavior. `MODE` selects HTTP, STDIO, or both; `PORT` sets the HTTP port; `DEFAULT_SEARCH_ENGINE` chooses the default engine; and `ALLOWED_SEARCH_ENGINES` limits the engines available to searches.

Set `USE_PROXY=true` and provide `PROXY_URL` to route runtime network traffic through an HTTP proxy. Installation proxy settings are separate from runtime proxy settings. `ENABLE_CORS` and `CORS_ORIGIN` configure cross-origin access. `SEARCH_MODE` selects request, automatic, or Playwright-based search behavior, with its documented effect currently applying to Bing.

## Tools and capabilities

- `search` queries the configured multi-engine search providers.
- `fetchLinuxDoArticle` retrieves a Linux.do post.
- `fetchCsdnArticle` retrieves a CSDN article.
- `fetchGithubReadme` retrieves a repository README from GitHub.
- `fetchWebContent` retrieves content from a public HTTP(S) URL.
- `fetchJuejinArticle` retrieves a Juejin post.

The server supports configurable result counts and HTTP proxy access for restricted resources. It can run in STDIO mode for MCP clients, HTTP mode, or both.

## Limitations and notes

Content extraction depends on the target site's accessibility and page structure. Homepages and JavaScript-heavy landing pages may not yield readable article text through generic fetching, so specific result pages are preferable. Playwright-based rendering cannot work unless Playwright and a usable browser are available. TLS verification can be disabled for the request leg of generic fetching with `FETCH_WEB_INSECURE_TLS`, but that setting does not change browser navigation behavior.

_Full upstream README: https://allmcps.com/mcp/aas-ee-open-websearch/readme_

