MCP Server My Lark Doc

Name: MCP Server My Lark Doc
Author: T0UGH

@T0UGH

リポジトリ

MCP Server My Lark Doc について

概要はまだありません

基本情報

カテゴリ

コミュニケーション

ライセンス

MIT

ランタイム

python

トランスポート

stdio

公開者

T0UGH

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "mcp_server_my_lark_doc": {
      "command": "uvx",
      "args": [
        "mcp-server-my-lark-doc"
      ]
    }
  }
}

ツール

Retrieve document content from Lark

documentUrl (string) - The URL of the Lark document

Document content in text format

Doc URLs: https://xxx.feishu.cn/docx/xxxxx

Search documents in Lark Wiki

query (string) - Search keywords

JSON string containing search results with following fields:

概要

What is MCP Server My Lark Doc?

A Model Context Protocol server for searching and accessing Lark (Feishu) documents. It supports both Lark Doc and Wiki document types, automatically handles document type detection and ID extraction, and returns raw content in text format for LLM processing. Designed for developers who need to integrate Lark document content into AI workflows.

How to use MCP Server My Lark Doc?

Install via uvx mcp-server-my-lark-doc. Set the environment variables LARK_APP_ID and LARK_APP_SECRET (from the Lark Open Platform), and optionally OAUTH_HOST and OAUTH_PORT. Then configure the server in Claude desktop's mcpServers block as shown in the README. Once connected, use the two available tools: get_lark_doc_content (accepts a document URL) and search_wiki (accepts a query and optional page size).

Key features of MCP Server My Lark Doc

OAuth-based user authentication with automatic token refresh
Supports both Lark Doc and Wiki document types
Automatic document type detection and ID extraction
Two tools: document content retrieval and wiki search
Comprehensive error reporting with clear troubleshooting messages
Customizable OAuth callback server host and port

Use cases of MCP Server My Lark Doc

Retrieving Lark document content for large language model (LLM) processing
Searching Lark Wiki for relevant documents by keyword
Automating document access in a Lark-integrated AI assistant
Integrating Lark document data into Claude desktop workflows

FAQ from MCP Server My Lark Doc

What permissions does my Lark app need?

Your Lark app must have the following permissions: wiki:wiki:readonly, wiki:node:read, docx:document:readonly, and search:docs:read.

How do I set up OAuth for this server?

Create a Lark application on the Lark Open Platform, obtain your App ID and App Secret, and export them as environment variables LARK_APP_ID and LARK_APP_SECRET. The server starts an OAuth callback server on localhost:9997 by default.

What document URL formats are supported?

The get_lark_doc_content tool accepts Doc URLs (e.g., https://xxx.feishu.cn/docx/xxxxx) and Wiki URLs (e.g., https://xxx.feishu.cn/wiki/xxxxx).

What happens if authorization times out?

If the user does not complete OAuth within 5 minutes, the server returns an "Authorization timeout" error.

Can I customize the OAuth callback server?

Yes. Use the OAUTH_HOST and OAUTH_PORT environment variables to change the callback server host and port from the defaults (localhost and 9997).