
IBM AS400 ISeries MCP Server
This project implements a customizable MCP server to execute RTE protocol. The intention of this project is to be easily extensible. You can add your own MCP tools by using RteUtils class. Also is needed to define the RTE connection parameters and the RTE flow in a yaml format.
README
IBM AS/400 ISeries MCP Server
This project implements a customizable MCP server to execute RTE protocol.
Example
Development
Requirements
- Java 21+*
Run
-
Compile project
./mvnw clean package
-
Run server on SSE mode
java -jar target/rte-mcp-server-0.0.1-SNAPSHOT.jar
-
Run server on STDIO mode
java -jar -Dspring.ai.mcp.server.stdio=true rte-mcp-server-0.0.1-SNAPSHOT.jar
RteUtils
The intention of this project is to be easily extensible. You can add your own MCP tools by
using RteUtils
class.
Also is needed to define the RTE connection parameters and the RTE flow in a yaml format.
RteUtils
has the following methods implemented:
static RteProtocolClient connect(RteConfig config)
: Connects to the RTE server using the provided yaml.static String executeSteps(RteProtocolClient client, List<RteStep> steps)
: Executes the provided steps on the RTE server and returns the screen result as a string using the provided yaml.
The method exposed as MCP tool should be annotated as @Tool
and the class should be annotated as @Services
like in the
followig example:
@Service
public class RteServiceLoginTool implements RteTool {
private static final Logger logger = LoggerFactory.getLogger(RteServiceLoginTool.class);
@Tool(description = "Execute an RTE transaction in a given server", name = "TN5250 Login")
public String executeRteTransaction(String server, int port) {
try {
RteFlow flow = RteYamlLoader.load("/flow.yaml");
RteProtocolClient client = RteTool.connect(flow.getConfig(), server, port);
return RteTool.executeSteps(client, flow.getSteps());
} catch (IOException | RteIOException | InterruptedException | TimeoutException e) {
logger.error("Error executing RTE flow", e);
return "Error executing RTE flow: " + e.getMessage();
}
}
}
The yaml defines the connection parameters and the steps to execute in the RTE server. This * yaml* file should have the following structure
config:
server: localhost
port: 2324
protocol: TN5250
terminalType: 'IBM-3477-FC: 27x132'
sslType: None
steps:
- actions:
- label: User
input: TESTUSR
- label: Password
input: TESTPSW
attentionKey: ENTER
Any number the steps can be defined in the yaml file.
-
server: The server address to connect to the RTE server.
-
port: The port to connect to the RTE server.
-
protocol: The supported protocols are TN5250, TN3270 and VT420
-
terminalType: The following table shows the supported terminal types for each protocol:
Protocol Terminal Type TN5250 IBM-3477-FC: 27x132 IBM-3179-2: 24x80 TN3270 IBM-3278-M2: 24x80 IBM-3278-M2-E: 24x80 IBM-3278-M3: 32x80 IBM-3278-M3-E: 32x80 IBM-3278-M4: 43x80 IBM-3278-M4-E: 43x80 IBM-3278-M5: 27x132 IBM-3278-M5-E: 27x132 VT420 VT420-7: 24x80 -
sslType: The supported ssl types are None TLS and SSLv3.
-
steps: A list of actions representing the steps to execute in the RTE server.
-
actions: All the actions to execute in the current screen, any number of pairs label/input could be added.
- label: The label of the field to fill in the RTE screen.
- input: The value to fill in the field.
- attentionKey: The key to press after filling the fields, it can be ENTER, F1, F2, F3, F4, F5, F6, F7, F8, F9, F10, F11, F12, F13, F14, F15, F16, F17, F18, F19, F20, F21, F22, F23, F24, ATTN, CLEAR, SYSRQ, RESET, ROLL_UP, ROLL_DN, PA1, PA2 or PA3.
Testing
An example of integration test is provided in the RteServiceLoginToolTest
class, this test uses
the MCPClient
validate the list of tools and execute it.
note: to test the behaving of to tool is recommended to generate a dump of the RTE flow to be executed and mock it using wiresham
More information on how to test and debug the server could be found here
推荐服务器

Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。