bitable
This MCP server provides access to Lark Bitable through the Model Context Protocol. It allows users to interact with Bitable tables using predefined tools.
claude mcp add --transport stdio lloydzhou-bitable-mcp uvx bitable-mcp \ --env APP_TOKEN="your-app-token" \ --env PERSONAL_BASE_TOKEN="your-personal-base-token"
How to use
This MCP server exposes Lark Bitable data via the Model Context Protocol, enabling Claude and other clients to interact with Bitable tables using predefined tools. The available tools let you list the tables, describe a table's schema, and execute read queries against the underlying Bitable data. Specifically, you can:
- list_table: Retrieve a JSON-encoded list of table names for the connected Bitable workspace.
- describe_table: Provide a table name to obtain a JSON-encoded list of its columns and types.
- read_query: Run a SQL-like query to read data from the tables and receive a JSON-encoded result set. To use these tools, ensure your environment is configured with a PERSONAL_BASE_TOKEN and APP_TOKEN, which authenticate your access to Bitable. The server can be run via uvx (as shown in the installation steps) and will expose these tools to your MCP-enabled client. If you prefer Python, you can also run the server with python -m bitable_mcp, which uses the same tool set through the same API surface.
How to install
Prerequisites:
- Install uvx in your environment (or use Python with the bitable_mcp module).
- Obtain your PERSONAL_BASE_TOKEN and APP_TOKEN from your Bitable account.
Option A: Using uvx (recommended for Claude CLI integration)
- Install uvx and start the server:
# Example: run the MCP server with uvx
PERSONAL_BASE_TOKEN=your_personal_base_token APP_TOKEN=your_app_token uv run --with uv --with bitable-mcp bitable-mcp-install
- Verify the server is running and accessible via MCP tooling.
Option B: Using pip installation (Python)
- Install the package:
pip install bitable-mcp
- Run via python -m bitable_mcp:
PERSONAL_BASE_TOKEN=your_personal_base_token APP_TOKEN=your_app_token python -m bitable_mcp
- Confirm the server is running and the tools are available through your MCP client.
Option C: Configure for Claude or Zed (as shown in the README)
- For Claude via uvx:
"mcpServers": {
"bitable-mcp": {
"command": "uvx",
"args": ["bitable-mcp"],
"env": {
"PERSONAL_BASE_TOKEN": "your-personal-base-token",
"APP_TOKEN": "your-app-token"
}
}
}
- For Claude via Python (module):
"mcpServers": {
"bitable-mcp": {
"command": "python",
"args": ["-m", "bitable_mcp"],
"env": {
"PERSONAL_BASE_TOKEN": "your-personal-base-token",
"APP_TOKEN": "your-app-token"
}
}
}
Include the same environment variables in your Zed settings if you configure through that client as well.
Additional notes
Tips and common considerations:
- Ensure your PERSONAL_BASE_TOKEN and APP_TOKEN are kept secret and not checked into version control.
- If you switch between uvx and Python, keep the tokens consistent across configurations.
- When troubleshooting, you can use the MCP inspector to debug uvx installations as noted in the README.
- If your environment uses a firewall or proxy, make sure MCP traffic to Bitable APIs is allowed.
- The current tooling supports list_table, describe_table, and read_query; newer tools may be added, so check for updates in the MCP server repository or README for changes.
Related MCP Servers
web-eval-agent
An MCP server that autonomously evaluates web applications.
mcp-neo4j
Neo4j Labs Model Context Protocol servers
Gitingest
mcp server for gitingest
zotero
Model Context Protocol (MCP) server for the Zotero API, in Python
fhir
FHIR MCP Server – helping you expose any FHIR Server or API as a MCP Server.
unitree-go2
The Unitree Go2 MCP Server is a server built on the MCP that enables users to control the Unitree Go2 robot using natural language commands interpreted by a LLM.