mcp -foundry

A Model Context Protocol server for interacting with Foundry

Installation

Run this command in your terminal to add the MCP server to Claude Code.

Run in terminal:

Command

claude mcp add --transport stdio qwert666-mcp-server-foundry uv --directory <path_to_mcp_server> run mcp-server-foundry \
  --env TOKEN="<token>" \
  --env SCOPES="<scopes>" \
  --env HOSTNAME="<hostname>" \
  --env CLIENT_ID="<client_id>" \
  --env ONTOLOGY_ID="<ontology_id>" \
  --env CLIENT_SECRET="<client_secret>"

How to use

This MCP server provides a Foundry integration for interacting with datasets, ontology objects, and functions within a Foundry instance. It exposes a set of tools to list and query datasets, list and query ontology objects, and to list and execute functions. You connect to Foundry by configuring the environment variables (hostname, token or OAuth2 flow with client credentials) and then launching the server via the uv command. The available tools allow AI assistants to discover data assets, inspect ontology objects, and trigger predefined functions inside your Foundry workspace.

How to install

Prerequisites:

Python 3.9+
Foundry Platform SDK (foundry-platform-sdk)
pyarrow
pandas
mcp (Model Context Protocol tooling)

Installation steps:

Install Python and dependencies
- Ensure Python 3.9+ is installed on your system.
- Install required Python packages: python -m pip install --upgrade pip python -m pip install pyarrow pandas foundry-platform-sdk
Prepare the MCP server code
- Clone the Foundry MCP server repository (or obtain the server package).
- Ensure the server directory contains the mcp-server-foundry code compatible with MCP.
Configure environment for Foundry access
- Set up the required environment variables (see mcp_config below for placeholders).
- If using OAuth flow, ensure CLIENT_ID, CLIENT_SECRET, and SCOPES are configured to obtain a token, or provide a TOKEN directly.
Run the server via MCP tooling
- Use the UV-based launcher as shown in the mcp_config example (below) to start the server:
uv --directory <path_to_mcp_server> run mcp-server-foundry

Note: You can alternatively adapt the commands to your environment or containerize the setup with Docker if preferred.

Additional notes

Environment variables are required for authentication with Foundry: HOSTNAME, TOKEN or CLIENT_ID/CLIENT_SECRET with SCOPES, and ONTOLOGY_ID. If TOKEN is not provided, the server will attempt an OAuth2 flow using CLIENT_ID, CLIENT_SECRET, and SCOPES. Ensure NETWORK access to your Foundry instance and that the ontology ID corresponds to a valid ontology in your Foundry workspace. Common issues include invalid HOSTNAME, expired or missing TOKEN, or insufficient SCOPES for the requested ontology interactions. When deploying in containers, remember to propagate environment variables into the container runtime.

Related MCP Servers

mcp-vegalite

MCP server from isaacwasserman/mcp-vegalite-server

github-chat

A Model Context Protocol (MCP) for analyzing and querying GitHub repositories using the GitHub Chat API.

nautex

MCP server for guiding Coding Agents via end-to-end requirements to implementation plan pipeline

pagerduty

PagerDuty's official local MCP (Model Context Protocol) server which provides tools to interact with your PagerDuty account directly from your MCP-enabled client.

futu-stock

mcp server for futuniuniu stock

mcp -boilerplate

Boilerplate using one of the 'better' ways to build MCP Servers. Written using FastMCP