Grist

Provides programmatic access to Grist data via MCP, enabling data exploration, editing, querying, and exports from language models.

python

29

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "nic01asfr-mcp-server-grist": {
      "command": "python",
      "args": [
        "-m",
        "grist_mcp_server"
      ],
      "env": {
        "LOG_LEVEL": "INFO",
        "GRIST_API_KEY": "votre_clé_api_grist",
        "GRIST_API_HOST": "https://docs.getgrist.com/api"
      }
    }
  }
}

You can run Grist's MCP server to access and manipulate Grist data directly from language models and other MCP clients. It exposes a programmable interface for listing organizations, workspaces, documents, tables, records, and more, with support for filtering, SQL-like queries, data export, and webhook integration. This guide walks you through practical use, installation steps, and key configuration details to get you up and running quickly.

How to use

Start by choosing how you want to run the Grist MCP server. You can operate it as a local stdio server that you invoke from your client or as a Docker-based service that runs in a container. Your MCP client will connect through the configured transport and send requests for reading, creating, updating, and deleting Grist data, as well as for filtering, joining, and exporting results.

How to install

Prerequisites you need before installation:

Python 3.8+
A valid Grist API key
A supported MCP client setup (your language environment or CLI)

Install the MCP server package with Python’s package manager, then run the server using one of the supported configurations.

pip install mcp-server-grist

After installation, you can run the server in one of the supported ways explained below.

## Configuration and start options

Configure and start the MCP server using one of the available transport methods. The server supports stdio, HTTP streamable, and Server-Sent Events transports. You can enable debugging to get detailed logs for troubleshooting.

For local development and Claude Desktop usage, you can define MCP connections like this in your client configuration. The following options show how to run Python, Node, or Docker-based setups.

## Available tools

### list\_organizations

List all organizations available in Grist, serving as a starting point to explore workspaces and documents.

### list\_workspaces

List workspaces within an organization to locate the target document set.

### list\_documents

List documents within a workspace for quick access to a specific document.

### list\_tables

List tables inside a document to inspect data structure.

### list\_columns

List columns in a table to understand the data schema.

### list\_records

List records from a table with sorting and limiting options.

### filter\_sql\_query

Run simple SQL-like filters on a single table to narrow results.

### execute\_sql\_query

Run complex SQL queries with joins, subqueries, and parameter support.

### add\_grist\_records

Add new records to a table.

### update\_grist\_records

Update existing records in a table.

### delete\_grist\_records

Delete records from a table.

### download\_document\_excel

Export a document to Excel for offline analysis.

### download\_table\_csv

Export a single table to CSV for data sharing.

### list\_webhooks

List webhooks configured for a document to integrate external services.

### create\_webhook

Create a new webhook to notify external systems on document events.