Overview and Installation Guide for Various MCP Services and Their Use with Sequential Thinking for Manus‑like Effects

1. Sequential Thinking MCP Description: An AI "deep thinking" mode that decomposes complex problems through multi‑turn reasoning to produce comprehensive solutions. URL: https://github.com/smithery-ai/reference-servers/tree/main/src/sequentialthinking

Installation commands: npx -y @modelcontextprotocol/server-sequential-thinking (Mac) npm install -g @modelcontextprotocol/server-sequential-thinking (Windows) node C:\Users\heiye\AppData\Roaming\npm\node_modules\@modelcontextprotocol\server-sequential-thinking\dist\index.js

2. Firecrawl MCP Description: Enterprise‑grade web data extraction tool supporting JavaScript rendering, batch processing, intelligent content search, and deep crawling. URL: https://github.com/mendableai/firecrawl-mcp-server

Installation (set environment variable FIRE_CRAWL_API_KEY or FIRECRAWL_API_KEY ) and run one of the following: npx -y mcp-server-firecrawl (Mac) npm install -g mcp-server-firecrawl && node C:\Users\heiye\AppData\Roaming\npm\node_modules\mcp-server-firecrawl\dist\src\index.js (Windows) Alternative commands: npx -y firecrawl-mcp , npm install -g firecrawl-mcp && node C:\Users\heiye\AppData\Roaming\npm\node_modules\firecrawl-mcp\dist\index.js , or cmd /c "set FIRECRAWL_API_KEY=YOUR_KEY && npx -y firecrawl-mcp" .

3. Fetch MCP Description: Anthropic‑developed tool for efficient web content scraping, returning markdown. URL: https://github.com/smithery-ai/mcp-fetch

Installation: npx -y @kazuph/mcp-fetch (Mac) npm install -g @kazuph/mcp-fetch && node C:\Users\heiye\AppData\Roaming\npm\node_modules\@kazuph\mcp-fetch\dist\index.js (Windows) Or use Python: pip install mcp-server-fetch python -m mcp_server_fetch uvx mcp-server-fetch

4. Hot News MCP Description: Retrieves hottest news from platforms such as Douyin, Zhihu, Tieba, Weibo, Baidu, Bilibili, and Douban. URL: https://github.com/wopal-cn/mcp-hotnews-server

Installation: npx -y @wopal/mcp-server-hotnews (Mac) npm install -y @wopal/mcp-server-hotnews && node C:\Users\heiye\AppData\Roaming\npm\node_modules\@wopal\mcp-server-hotnews\build\index.js (Windows)

5. Playwright MCP Description: Powerful automation testing tool for end‑to‑end tests across multiple browsers. URL: https://github.com/executeautomation/mcp-playwright

Installation: npx -y @executeautomation/playwright-mcp-server (Mac) npm install -g @executeautomation/playwright-mcp-server && node C:\Users\heiye\AppData\Roaming\npm\node_modules\@executeautomation\playwright-mcp-server\dist\index.js (Windows)

6. Magic MCP Description: Enables rapid creation of front‑end UI components from 21st.dev via the /ui command. URL: https://github.com/21st-dev/magic-mcp

7. Brave Search MCP Description: Uses Brave Search API for information retrieval (requires API key). URL: https://github.com/smithery-ai/reference-servers/tree/main/src/brave-search

Installation: npx -y @modelcontextprotocol/server-brave-search (Mac) npm install -g @modelcontextprotocol/server-brave-search && node C:\Users\heiye\AppData\Roaming\npm\node_modules\@modelcontextprotocol\server-brave-search\dist\index.js (Windows)

Installation Tips On Windows, you can test the service with the node command; replace the placeholder username heiye with your own.

Implementing a Manus‑like Effect Core idea: Combine a Sequential Thinking prompt with the Cursor AI agent to emulate the complex behavior of the Manus agent. The workflow consists of four steps:

Set up the required MCP services.

Submit the task description.

Interact with the agent during execution.

Review the final generated web page.

Images illustrating each step are included in the original article.

Simple Summary MCP services offer great imagination potential, especially when integrated with large models or existing data/services. Traditional companies can transform into MCP providers (e.g., Brave Search for search, Firecrawl for crawling). By coupling prompts with MCP services, complex agent scenarios like Manus can be built quickly, though foundational AI agents such as Cursor or Windsurf are still needed.