This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
🤖 Feishu MCP Server — Let AI Agents directly control Feishu: send messages, create calendars, manage tasks, and write cloud documents.
Related MCP server: Feishu MCP Server
This project implements an MCP (Model Context Protocol) Server that wraps Feishu’s core capabilities as structured AI tools, enabling any MCP-compatible AI Agent / LLM toolchain to:
Capability | Functions |
💬 Messages | Send & reply to group messages, @mention members, Markdown format |
📅 Calendar | Create/query group calendars, create events, invite attendees |
✅ Tasks | Create tasks, assign owners, set due dates |
📄 Documents | Create cloud docs, write Markdown content, upload files, one-click share links, set collaborator permissions |
👥 Users | Get group member lists, resolve users by name |
🔔 Events | Receive Feishu push events via long-connection WebSocket in real time (no public IP required) |
git clone https://github.com/your-username/feishu-miqroera-mcp.git
cd feishu-miqroera-mcp
python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS / Linux
source .venv/bin/activate
pip install -e .cp .env.example .envEdit .env:
FEISHU_APP_ID=cli_xxxxxxxxxxxxxxxxxx
FEISHU_APP_SECRET=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxHow to get credentials: log in to the Feishu Open Platform → create an in-house app → go to the "Credentials & Basic Info" page
# Confirm the MCP server starts correctly (Ctrl+C to exit)
.venv\Scripts\python.exe -m feishu_mcp.serverYou should see MCP server running on stdio indicating success.
NanoBot is a lightweight multimodal AI Agent framework with native support for MCP Servers and Feishu channels.
Step 1: Edit the NanoBot config file
# Config file location (auto-created)
~/.nanobot/config.json # macOS / Linux
%USERPROFILE%\.nanobot\config.json # WindowsStep 2: Add feishu-mcp to tools.mcpServers
{
"tools": {
"mcpServers": {
"feishu-mcp": {
"command": "C:/path/to/feishu-miqroera-mcp/.venv/Scripts/python.exe",
"args": ["-m", "feishu_mcp.server"],
"env": {
"FEISHU_APP_ID": "cli_xxxxxxxxxxxxxxxxxx",
"FEISHU_APP_SECRET": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
}
}
},
"channels": {
"feishu": {
"appId": "cli_xxxxxxxxxxxxxxxxxx",
"appSecret": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
}
}Windows note: Use forward slashes
/or double backslashes\\in paths
Step 3: Start NanoBot
nanobot runNanoBot will automatically establish a Feishu long connection; the AI can then drive Feishu operations via natural language.
Step 1: Locate the config file
OS | Path |
Windows |
|
macOS |
|
Step 2: Add the MCP Server config
{
"mcpServers": {
"feishu-mcp": {
"command": "C:/path/to/feishu-miqroera-mcp/.venv/Scripts/python.exe",
"args": ["-m", "feishu_mcp.server"],
"env": {
"FEISHU_APP_ID": "cli_xxxxxxxxxxxxxxxxxx",
"FEISHU_APP_SECRET": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
}
}
}macOS example:
"command": "/Users/yourname/feishu-miqroera-mcp/.venv/bin/python"
Step 3: Restart Claude Desktop
After restarting, look for the 🔧 icon in the chat input area and confirm "feishu-mcp" is loaded.
Step 1: Locate the config file
<project root>/.cursor/mcp.jsonOr global config:
OS | Path |
Windows |
|
macOS |
|
Step 2: Add config
{
"mcpServers": {
"feishu-mcp": {
"command": "C:/path/to/feishu-miqroera-mcp/.venv/Scripts/python.exe",
"args": ["-m", "feishu_mcp.server"],
"env": {
"FEISHU_APP_ID": "cli_xxxxxxxxxxxxxxxxxx",
"FEISHU_APP_SECRET": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
}
}
}Step 3: Open Cursor → Settings → MCP → Confirm feishu-mcp status is green
All tools that support MCP stdio transport can connect using the same config format:
{
"command": "/absolute/path/to/.venv/bin/python",
"args": ["-m", "feishu_mcp.server"],
"env": {
"FEISHU_APP_ID": "cli_xxx",
"FEISHU_APP_SECRET": "xxx"
}
}Feishu's long connection (WebSocket) requires no public IP and no ngrok — receive Feishu push events directly from behind NAT.
# Start the event listener standalone (runs 24/7)
.venv\Scripts\python.exe -m feishu_mcp.webhook.longconnExample log output after startup:
INFO Connecting to Feishu WebSocket: wss://msg-frontier.feishu.cn/ws/v2
INFO Feishu long connection established
INFO Received message event: chat_id=oc_xxx, sender=ou_xxx, text=HelloReceiving events in your own application code:
import lark_oapi as lark
client = lark.Client.builder() \
.app_id("cli_xxx") \
.app_secret("xxx") \
.event_callback(lark.EventType.IM_MESSAGE_RECEIVE_V1, your_handler) \
.build()
ws = lark.ws.Client(app_id, app_secret, event_handler=client.event_handler)
ws.start()# Build image
docker build -t feishu-miqroera-mcp .
# Start Feishu event listener (background)
docker run -d \
--name feishu-listener \
--env-file .env \
--restart unless-stopped \
feishu-miqroera-mcp \
feishu_mcp.webhook.longconn
# View logs
docker logs -f feishu-listener# Copy and fill in env vars
cp .env.example .env
# Edit .env with APP_ID and APP_SECRET
# Start event listener
docker compose up -d feishu-listener
# Check status
docker compose ps
# View logs
docker compose logs -f feishu-listenerMCP Server (stdio) is usually invoked directly as a local process by the AI Agent framework and does not need to be containerized.
Docker is primarily used for persistently running the event listener.
Tool | Description |
| Get the member list of a group |
| Look up users in a group by name |
| Send a text/rich-text message to a group or user |
| Reply to a specific message |
| Get or create a shared group calendar |
| Create an event in a calendar |
| Add attendees to a calendar event |
| Query the list of calendar events |
| Create a Feishu task (with due time and description) |
| Assign an owner to a task |
| Add a task to a tasklist |
| Query the task list |
| Create a folder in Drive |
| Create a Feishu cloud document |
| Write Markdown content into a document (accepts docx |
| Upload a file to Feishu Drive |
| Upload a file and return a shareable link in one step (upload + set permission + get link) |
| Insert a file attachment block into a document (accepts docx |
| Add collaborators to a document (supports users or groups) |
| Set document public access / link sharing permission |
| Get the sharing link for a document |
| Handle a permission request and authorize the applicant |
See docs/api.md for full parameter details.
Enable the following permissions in your app's management page on the Feishu Developer Console:
Permission | Purpose |
| Send/receive messages |
| Group @mention feature |
| Read group members |
| Task read/write |
| Calendar read/write |
| Drive/document read/write |
| Cloud document content editing |
| Resolve wiki node tokens to docx document IDs (required if documents are wiki-mounted) |
Path to enable permissions: App Management → Permission Management → Enable the above permissions → Publish version.
# Run unit tests
pytest tests/ -v
# Run integration tests (requires real Feishu credentials)
pytest tests/integration/ -v
# Format code
ruff format src/ tests/
# Lint check
ruff check src/ tests/See CONTRIBUTING.md for details.
MIT © 2026 feishu-miqroera-mcp contributors
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Messaging tools for AI agents: send messages, manage chats, groups and channels.
Connect AI assistants to Google Sheets through controlled tools for reading and updating rows.
Calendar API for AI agents: events, availability, Google/Microsoft setup, scheduling, and iCal.
Connect your AI agent to 300+ methods and 100+ live tools
