Wordware MCP Server
Integrates Wordware deployed flows as tools usable within Claude conversations via the Model Context Protocol.
README
Wordware MCP Server
A Model Context Protocol (MCP) server implementation that enables integration of your Wordware deployed flows as tools that can be used directly within Claude conversations.
Features
- Integration with Claude via MCP
- Support for basic Wordware tools including:
- Research Founder: Comprehensive founder analysis and meeting prep
- Lead Enrichment: Sales prospect research and intelligence
- Save to Notion: Direct page saving to Notion
- React Agent: Task solving with Google search and API capabilities
Setup
1. Modify the env.example.ts file in the root directory to set it to env.ts and set the variables inside:
Using a regular .env file seemed to complicated considering you would also need to make those variables accessible from the server's environment (inside the Claude Desktop config). Come back to this if you find a better solution.
OPENAI_API_KEY: Your OpenAI API key (used inadd-tool.tsto generate a zod schema from the given information about the Wordware flow)NOTION_SECRET: Your Notion secretNOTION_PARENT_PAGE_ID: The ID of the Notion page to save toSAVE_TO_NOTION_APP_ID: The ID of your deployed Save to Notion appRESEARCH_FOUNDER_APP_ID: The ID of your deployed Research Founder appLEAD_ENRICHMENT_APP_ID: The ID of your deployed Lead Enrichment appREACT_AGENT_APP_ID: The ID of your deployed React Agent app
To setup Notion for the saveToNotion tool, follow the instructions here
If you don't want to use one of the 4 tools, you can just leave its corresponding env variable empty and the tool will not be registered.
2. Install dependencies and build the server:
npm install
npm run build-server
3. Testing the server with Claude for Desktop
You need to have Claude for Desktop installed to test the server. If you do, you need to modify the config file to use the MCP server. Run the following command to open the config file (if you use VSCode):
code ~/Library/Application\ Support/Claude/claude_desktop_config.json
Then, add the following to the file (make sure to replace /ABSOLUTE/PATH/TO/PARENT/FOLDER/wordware/build/index.js with the absolute path to the index.js file in the build folder of this repository):
{
"mcpServers": {
"wordware": {
"command": "node",
"args": ["/ABSOLUTE/PATH/TO/PARENT/FOLDER/wordware/build/index.js"]
}
}
}
4. Using the add-tool command (optional but recommended)
Note: The add-tool command has only been tested on a few simple Wordware flows. Since it relies on OpenAI to generate the schema, please make sure to verify the output of the tool before using it.
To run the tool:
npm run add-tool
Future Improvements
Dynamic Tool Configuration
Currently, the tools are hardcoded in the server implementation. Future improvements should focus on:
- Implement authentication with Wordware API
- Add capability to fetch user's deployments automatically
- Figure out which deployments can be added as MCP tools automatically
- Better support for complex blocks (e.g. "ask a Human" blocks)
- Better error handling for tool calls
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。