Tetra Data Assistant MCP Server User Guide (Beta)

Learn how to install, configure, and use the Tetra Data Assistant MCP Server with your AI coding assistant.

This guide shows how to install, configure, and use the Tetra Data Assistant MCP Server. For an overview of available tools and prompts, see the Tetra Data Assistant MCP Server overview page.

📘
NOTE
The Tetra Data Assistant MCP Server is currently available in beta release. For more information, or to get access, contact your customer account leader.

🚧
IMPORTANT
Code and artifacts generated by AI agents using this MCP server should always be reviewed for accuracy, data integrity, and security before use — as with any AI-generated code.

Prerequisites

To install and run the Tetra Data Assistant MCP Server, you need the following:

Python 3.11 or later. To check your Python version, run python3 --version.
A Python CLI tool installer — either of the following:
- pipx — Install by running brew install pipx (macOS) or pip install pipx (all platforms). For other methods, see the pipx installation guide.
- uv — A fast Python package installer and resolver. Install by running brew install uv (macOS) or pip install uv (all platforms). For other methods, see the uv installation guide.
Access to the TetraScience package repository. The MCP server isn't published to the public Python Package Index (PyPI). It's distributed from a private TetraScience package repository, and installing it requires a username and reference token issued by TetraScience. To request credentials, contact your customer account leader. For details, see Get Access to the TetraScience Package Repository.

Get Access to the TetraScience Package Repository

TetraScience distributes the MCP server from the ts-pypi-external repository at https://tetrascience.jfrog.io/artifactory/api/pypi/ts-pypi-external/simple. TetraScience issues one set of credentials per customer organization:

Username — The repository username that TetraScience provides with your reference token.
Reference token — A token that acts as the password. Reference tokens expire one year after they're issued. If installation fails with an authentication error, contact your customer account leader to request a new token.

🚧
IMPORTANT
Treat your reference token as a secret. Don't commit it to source control, paste it into shared configuration files, or share it outside your organization.

How you use these credentials depends on how your organization manages Python packages:

If your organization runs its own package repository (for example, JFrog Artifactory or Sonatype Nexus), TetraScience recommends that your repository administrator add ts-pypi-external as a remote repository and include it in the virtual (group) repository that your developers already install Python packages from. Your developers then install the MCP server from your internal repository without handling TetraScience credentials themselves. Make sure your repository resolves ts-mcp-tetra-data-assistant from ts-pypi-external only, and never from public PyPI.
If your developers install packages directly, they configure the credentials on their own machine. Use the Install by Using uv steps.

Install the MCP Server

📘
NOTE
Install a v0.5.x release of the MCP server. Version 0.6.0 and later are designed to work with a TetraScience plugin that isn't yet available to customers. The commands in this section install the latest v0.5.x release.

Install by Using uv

To install the MCP server directly from the TetraScience package repository by using uv, do the following. Commands are shown for macOS and Linux shells (bash or zsh) and for Windows PowerShell. Select the tab for your shell. On Windows, use PowerShell, not Command Prompt.

Set your repository username as an environment variable, replacing the placeholder with the username that TetraScience issued to your organization:

export UV_INDEX_TETRASCIENCE_USERNAME="<username>"

$env:UV_INDEX_TETRASCIENCE_USERNAME = "<username>"

Run the following command, then paste your reference token and press Return. The token is hidden as you paste it, and it isn't saved in your shell history:

read -rs UV_INDEX_TETRASCIENCE_PASSWORD && export UV_INDEX_TETRASCIENCE_PASSWORD

$env:UV_INDEX_TETRASCIENCE_PASSWORD = [System.Net.NetworkCredential]::new("", (Read-Host -AsSecureString "Reference token")).Password

Run the following command. It's the same in every shell:

uv tool install --index tetrascience=https://tetrascience.jfrog.io/artifactory/api/pypi/ts-pypi-external/simple "ts-mcp-tetra-data-assistant<0.6"

This installs the server and makes the mcp-tetra-data-assistant command available globally. uv installs the MCP server from the TetraScience package repository and installs its open-source dependencies from public PyPI.

📘
NOTE
By default, uv installs each package from the first index that provides it, and doesn't check later indexes for that package. Because the TetraScience repository is listed first and provides ts-mcp-tetra-data-assistant, uv installs it from there instead of from public PyPI. If the installation fails with an authentication error, fix your credentials and run the same command again. Don't remove the --index option to work around the error. For more information, see Package indexes in the uv documentation.

📘
NOTE
If uv warns that a directory is not on your PATH, the mcp-tetra-data-assistant command won't be found until you add it. Run uv tool update-shell, then open a new terminal window.

After the installation succeeds, remove your credentials from the session:

unset UV_INDEX_TETRASCIENCE_USERNAME UV_INDEX_TETRASCIENCE_PASSWORD

Remove-Item Env:UV_INDEX_TETRASCIENCE_USERNAME, Env:UV_INDEX_TETRASCIENCE_PASSWORD

The credentials are only needed to download the package. Running the MCP server doesn't require them.

Install by Using pipx

Use pipx if your organization's package repository already provides ts-mcp-tetra-data-assistant (see Get Access to the TetraScience Package Repository). Run the following command, replacing <repository-url> with the URL of your organization's virtual (group) Python package index. This must be an index that serves both TetraScience packages and public PyPI packages, because the MCP server's open-source dependencies come from PyPI. A URL that only proxies ts-pypi-external won't work.

pipx install --index-url <repository-url> "ts-mcp-tetra-data-assistant<0.6"

If pip on your machine is already configured to use your organization's repository (for example, through pip.conf or the PIP_INDEX_URL environment variable), you can omit --index-url. To find your repository URL, contact your organization's repository administrator.

This installs the server and makes the mcp-tetra-data-assistant command available globally.

Configure Your AI Coding Assistant

After you install the MCP server, configure your AI coding assistant to use it. The following sections provide instructions for each supported assistant.

For a full guide on configuring custom MCP servers, see the VS Code documentation.

To see all available command-line options, run:

mcp-tetra-data-assistant --help

Find the MCP Server Command Path

For Claude Desktop, GitHub Copilot, and Cursor, use the full path to the installed command in the JSON configuration. To find it, run the command for your shell:

which mcp-tetra-data-assistant

(Get-Command mcp-tetra-data-assistant -CommandType Application).Path

Replace /full/path/to/mcp-tetra-data-assistant in the JSON examples below with the returned path, including the .exe extension on Windows. In JSON, escape each Windows backslash as \\. For example, if the returned path is C:\Users\yourname\.local\bin\mcp-tetra-data-assistant.exe, use:

{
  "command": "C:\\Users\\yourname\\.local\\bin\\mcp-tetra-data-assistant.exe"
}

Use your actual path in place of this example.

Augment

To add the Tetra Data Assistant MCP Server to Augment, do the following:

Install the MCP server as described in the Install the MCP Server section.
In VS Code, go to Augment Settings > Tools and click the Import from JSON button under MCP. Then, paste the following:

{
  "mcpServers": {
    "Tetra Data Assistant": {
      "command": "mcp-tetra-data-assistant",
      "args": []
    }
  }
}

Click Import. The MCP server starts and is available in Augment.

Alternatively, you can manually add the following to your Augment settings.json file:

{
  "augment.advanced": {
    "mcpServers": [
      {
        "name": "Tetra Data Assistant",
        "command": "mcp-tetra-data-assistant"
      }
    ]
  }
}

Claude Code

To add the Tetra Data Assistant MCP Server to Claude Code, do the following:

Install the MCP server as described in the Install the MCP Server section.
Run the following command to register the server globally (available in all projects):

claude mcp add --scope user "mcp-tetra-data-assistant" -- mcp-tetra-data-assistant

To register it for a single project only (creates or updates .mcp.json in the current directory), run:

claude mcp add --scope project "mcp-tetra-data-assistant" -- mcp-tetra-data-assistant

For more information, see the Claude Code documentation.

Claude Desktop

To add the Tetra Data Assistant MCP Server to Claude Desktop, do the following:

Install the MCP server as described in the Install the MCP Server section.
Find the full path to the installed command by following Find the MCP Server Command Path, including the Windows JSON escaping instructions if applicable.
In Claude Desktop, go to Settings, then select Developer under Desktop Settings.
Click the Edit Config button, which opens the claude_desktop_config.json file in your file explorer.
Add the following entry under mcpServers, replacing the placeholder with the full path to the mcp-tetra-data-assistant command from step 2:

{
  "mcpServers": {
    "mcp-tetra-data-assistant": {
      "command": "/full/path/to/mcp-tetra-data-assistant",
      "args": []
    }
  }
}

Save the file and restart Claude Desktop. The Tetra Data Assistant tools are available in your conversations.

GitHub Copilot

To add the Tetra Data Assistant MCP Server to GitHub Copilot in VS Code, do the following:

Install the MCP server as described in the Install the MCP Server section.
Verify that you have VS Code 1.99 or later and the GitHub Copilot extension installed.
Find the full path to the installed command by following Find the MCP Server Command Path, including the Windows JSON escaping instructions if applicable.
Create or edit .vscode/mcp.json in your workspace root and add the following, replacing the placeholder with the full path from the previous step:

{
  "servers": {
    "mcp-tetra-data-assistant": {
      "type": "stdio",
      "command": "/full/path/to/mcp-tetra-data-assistant",
      "args": []
    }
  }
}

Cursor

To add the Tetra Data Assistant MCP Server to Cursor, do the following:

Install the MCP server as described in the Install the MCP Server section.
Find the full path to the installed command by following Find the MCP Server Command Path, including the Windows JSON escaping instructions if applicable.
In Cursor, go to Settings > Cursor Settings > MCP.
Click Add new global MCP server and add the following, replacing the placeholder with the full path from step 2:

{
  "mcpServers": {
    "mcp-tetra-data-assistant": {
      "command": "/full/path/to/mcp-tetra-data-assistant",
      "args": []
    }
  }
}

For more information, see the Cursor MCP documentation.

Available Tools and Prompts

For a complete list of available tools and prompts, see the Tetra Data Assistant MCP Server overview page.

IDS Validator Configuration

The IDSValidator tool uses the ts-ids-validator tool to check compliance with IDS standards. To enable validating IDS compatibility with previous versions from TDP, configure the MCP server with the path to ts-sdk-cfg.json by using the --config flag. If not configured, the validator uses git tags.

For more information, see the ts-ids-validator README.

Limitations

The following are known limitations of the Tetra Data Assistant MCP Server:

The MCP server requires a Python CLI tool installer (pipx or uv) to install and run.
The MCP server isn't available on public PyPI. Installing it requires TetraScience package repository credentials.
AI-generated code and artifacts must be reviewed for accuracy, data integrity, and security before use.

Documentation Feedback

Do you have questions about our documentation or suggestions for how we can improve it? Start a discussion in TetraConnect Hub. For access, see Access the TetraConnect Hub.

📘
NOTE
Feedback isn't part of the official TetraScience product documentation. TetraScience doesn't warrant or make any guarantees about the feedback provided, including its accuracy, relevance, or reliability. All feedback is subject to the terms set forth in the TetraConnect Hub Community Guidelines.

Updated 2 days ago

Did this page help you?