jcrawl4ai-mcp-server
@rainyheart
Java implementation of MCP Server for Crawl4ai
Overview
What is jcrawl4ai-mcp-server?
jcrawl4ai-mcp-server is a Spring Boot-based MCP server (Java implementation) that interacts with the Crawl4ai API to perform web crawling. It allows users to crawl specified URLs with configurable strategy, maximum depth, and output format, and to retrieve crawl results by task ID.
How to use jcrawl4ai-mcp-server?
Configure the cawl4ai.base-url and cawl4ai.api-token properties in src/main/resources/application.properties. Build with Maven (mvn clean install) and run the resulting JAR (java -jar target/jcrawl4ai-mcp-server-1.0.0.jar). A pre-built JAR is also available for download. The server uses stdio transport as shown in the MCP configuration example.
Key features of jcrawl4ai-mcp-server
- Crawls URLs with strategy, max depth, and output format.
- Retrieves crawl results by task ID.
- Built with Spring Boot and Spring AI MCP Server.
- Offers two auto-approvable tools: "crawl" and "task".
- Communicates via stdio transport.
Use cases of jcrawl4ai-mcp-server
- Automating web scraping for data ingestion into AI workflows.
- Integrating structured web crawling into MCP-compatible clients.
- Monitoring web page content changes with controlled depth.
- Fetching crawl results asynchronously using task IDs.
FAQ from jcrawl4ai-mcp-server
What dependencies does jcrawl4ai-mcp-server require?
The project depends on Spring AI MCP Server, Spring Boot, and Hutool.
How do I configure the Crawl4ai server connection?
Set cawl4ai.base-url (e.g., http://your-crawl4ai-server-url:11235) and cawl4ai.api-token (your API token) in the application.properties file.
What transport does the server use?
The server uses stdio transport, as shown in the MCP configuration JSON with "transportType": "stdio".
Is there a pre-built JAR available?
Yes, a pre-built JAR (jcrawl4ai-mcp-server-1.0.0.jar) can be downloaded from the link provided in the README.
How do I get the crawl result for a completed task?
Use the task method with the task ID to retrieve the crawl result as a JSON string.
