Teable

You can connect Teable to large language models through the Teable MCP Server. It acts as a bridge that lets AI assistants query Teable data, explore schema structures, and retrieve records using natural language, enabling dynamic, data-aware conversations with your Teable instance.

How to use

You use an MCP client to start and interact with the Teable MCP Server. The server runs locally as a standard input/output (stdio) process and exposes tools to query data, fetch records, view schemas, and inspect histories. Start the server, then configure your MCP client to point to the local process. Once running, you can ask your AI to perform actions such as querying a table with filters, fetching a specific record by ID, retrieving a table’s field definitions, and listing available spaces, bases, tables, and views.

Practical patterns you can try include: querying a table with filters and sorts, requesting the latest records with a limit, exploring a base’s tables, and inspecting a table’s fields to understand data types. You can also retrieve the full change history of a specific record to observe its evolution over time.

How to install

Prerequisites: you need Node.js installed on your machine, and you will run the MCP server locally after building the project.

1. Install dependencies for the Teable MCP Server project in your workspace.

2. Build the project to generate the runtime files.

3. Run the MCP server using the stdio configuration shown in the code block. This starts a local node process that serves the MCP endpoints.

4. Ensure your environment variables are set so the server can authenticate with Teable and know where to reach the API.

Configuration and running details

Environment variables you must set for the Teable MCP Server are the Teable API key and the base API URL. You provide these values to the server at runtime so it can access your Teable instance.

{
  "mcpServers": {
    "teable": {
      "command": "node",
      "args": [
        "/absolute/path/to/teable-mcp-server/dist/index.js"
      ],
      "env": {
        "TEABLE_API_KEY": "your_teable_api_key",
        "TEABLE_BASE_URL": "https://app.teable.ai/api"
      }
    }
  }
}

Additional usage notes

If you are integrating with client tools like Claude Desktop or Cursor, configure the MCP server to run as a local stdio process and provide the required environment variables. The server will expose endpoints that let the AI agents perform the supported operations against your Teable data.

Features and endpoints

You gain access to a comprehensive set of endpoints to interact with Teable data. Available actions include querying data from a specific table with filtering, sorting, and limiting; retrieving a single record by ID; accessing a record’s history; listing spaces, bases, tables, and views; and fetching the full schema of a table.

Security and access control

Use a Teable API key with the least-privilege permissions necessary for the tasks you perform. Bind the key to the exact bases and tables needed by the MCP server, and rotate keys regularly following your security practices.

Available tools

query_teable

Query data from a specific table with filtering, sorting, and limiting. Supports views to filter results by defined database perspectives.

get_record

Retrieve details for a single record by its ID.

get_record_history

Access the full change history of a specific record to track its evolution over time.

list_spaces

List all spaces available to the user.

list_bases

List all bases within a specific space.

list_tables

List all tables within a specific base.

list_views

Retrieve all views within a table to understand different data perspectives.

get_table_fields

Fetch the full schema (field definitions) of a table.