Click on "Install 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., "@lark-mcpsearch the wiki for the 2024 product roadmap"
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.
lark-mcp
A Model Context Protocol (MCP) server for Feishu/Lark. Lets AI assistants (Cursor, Claude Desktop, etc.) search wiki pages and read/write Feishu documents on your behalf via OAuth.
Features
search_wiki — Search documents in Feishu Wiki by keywords
get_doc_content — Read a Feishu document or wiki page by URL
create_doc — Create a new Feishu document (with optional Markdown content)
update_doc — Replace or append content in an existing Feishu document (Markdown → Feishu blocks)
Requirements
Node.js >= 18.14.1
A Feishu open-platform app with the following permissions:
docx:document,docx:document:readonlywiki:wiki:readonlysearch:docs:read
Setup
Install dependencies:
npm installCopy the example environment file and fill in your credentials:
cp .env.example .envStart the MCP server:
node index.mjs
Environment Variables
Variable | Required | Description |
| Yes | Feishu app ID |
| Yes | Feishu app secret |
| No | Your Feishu tenant domain (default: |
| No | Local port for OAuth callback (default: |
| No | OAuth scopes (default: |
Cursor / Claude Desktop Integration
Add the following to your MCP config (e.g. ~/.cursor/mcp.json):
{
"mcpServers": {
"lark": {
"command": "node",
"args": ["/path/to/lark-mcp/index.mjs"],
"env": {
"LARK_APP_ID": "your_app_id",
"LARK_APP_SECRET": "your_app_secret",
"FEISHU_DOMAIN": "your-tenant.feishu.cn"
}
}
}
}License
MIT
This server cannot be installed
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.