feishu-mcp-server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@feishu-mcp-server read the content of document https://xxx.feishu.cn/docs/abcDEF123"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
飞书文档 MCP 服务
一个用于读取飞书文档内容的 MCP (Model Context Protocol) 服务。
功能
read_feishu_document: 读取飞书文档的完整内容get_feishu_document_meta: 获取文档元信息(标题)支持普通文档和知识库(Wiki)文档
Related MCP server: MCP Server My Lark Doc
前置要求
飞书自建应用 credentials
已开通文档读取权限
Node.js >= 18
安装
npm install
npm run build配置
复制
.env.example为.env填写你的飞书应用凭证:
FEISHU_APP_ID=your_app_id
FEISHU_APP_SECRET=your_app_secret飞书开放平台配置
创建自建应用:https://open.feishu.cn/
获取 App ID 和 App Secret
开通权限:
docx:document:readonlydocx:document:readwiki:wiki:readonly(知识库文档需要)
发布应用(或使用调试模式)
使用方式
直接运行
# 设置环境变量
export FEISHU_APP_ID=your_app_id
export FEISHU_APP_SECRET=your_app_secret
# 运行
npm startClaude Code 配置
在 claude_desktop_config.json 中添加:
{
"mcpServers": {
"feishu": {
"command": "node",
"args": ["/path/to/feishu-mcp-server/dist/index.js"],
"env": {
"FEISHU_APP_ID": "your_app_id",
"FEISHU_APP_SECRET": "your_app_secret"
}
}
}
}文档 ID 获取
支持三种输入方式:
1. 普通文档
https://xxx.feishu.cn/docs/abcDEF123文档 ID:abcDEF123
2. 知识库文档(Wiki)
https://xxx.feishu.cn/wiki/Bj3Jw2kfji45nhkK5MQcv3NynWe直接传入完整 URL 或 Wiki 节点 ID:Bj3Jw2kfji45nhkK5MQcv3NynWe
3. 直接传入 ID
工具会自动识别是普通文档还是 Wiki 节点,并自动解析。
Available Tools
2 toolsget_feishu_document_metaA
获取飞书文档的元信息(标题)。根据文档 ID 或完整 URL 获取文档标题,不返回文档内容。支持普通文档和知识库(Wiki)文档。
Inputs: document_id (required).
read_feishu_documentA
读取飞书文档的内容。根据文档 ID 或完整 URL 获取文档的标题和完整文本内容。支持普通文档和知识库(Wiki)文档。
Inputs: document_id (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 2 tools
read_feishu_document returns full content, while get_feishu_document_meta returns only metadata. Their purposes are clear and non-overlapping, so agents can easily select the right tool.
Both tools follow a consistent [verb]_feishu_document pattern, with read and get as action prefixes. The naming is uniform and predictable.
At 2 tools, the server feels thin for a document-focused MCP. While the current tools are focused, the small set suggests limited scope.
The server covers reading content and metadata but lacks listing, searching, creating, updating, or deleting documents. This is a significant gap for a Feishu document server, limiting agents to only read operations.
Maintenance
Related MCP Connectors
MCP-native open-source Notion alternative: read & write pages, databases and kanban boards.
Read and write your team's shared, AI-readable wiki from any MCP client.
Team docs served to AI agents over MCP - search, Markdown reads, version pinning, read audit.
MCP-native collaborative markdown editor with real-time AI document editing
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceAn MCP server that enables AI assistants to search, read, create, and update Feishu/Lark documents and wiki pages. It facilitates seamless document management and wiki searches through OAuth-authenticated API access.MIT
- AlicenseCqualityDmaintenanceEnables searching and retrieving content from Lark (Feishu) documents via MCP tools, supporting Doc and Wiki types with OAuth authentication.24MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI applications to access Feishu (Lark) knowledge base and cloud documents through the MCP protocol.26 npm1ISC
- FlicenseAqualityCmaintenanceEnables reading, creating, updating, and appending Feishu documents, as well as querying and updating Bitable fields and records via the MCP protocol.12-