mcp -data-wrangler

The mcp-data-wrangler server exposes data wrangling capabilities as part of the MCP ecosystem. It provides operations for data aggregation and descriptive statistics to help preprocess and analyze datasets within your model-context workflows. You can run the server locally using uv (Python) and connect to it via MCP tooling or the inspector, enabling you to test and debug your data wrangling pipelines in a familiar MCP environment. The repository’s example shows how to register the tool as an MCP server under the name data-wrangler and how to launch it with uv, pointing to the server’s source directory.

To use the server, start it with the uv command (or run the inspector and connect to the server). Once running, you can send model-context prompts that request aggregation, summary statistics, or transformations; the server will process the data according to the implemented wrangling operations and return structured results suitable for downstream MCP steps.

Prerequisites:

• Python 3.8+ and a functional Python environment
• pip and optional virtual environment tooling
• Git to clone the repository

Step-by-step installation:

1. Clone the repository:

git clone https://github.com/yourusername/mcp-data-wrangler.git
cd mcp-data-wrangler

2. Create and activate a virtual environment (recommended):

python -m venv .venv
# macOS/Linux
source .venv/bin/activate
# Windows
.venv\Scripts\activate

3. Install dependencies (development-friendly install):

pip install -e ".[dev]"

4. Run tests (optional, to verify setup):

pytest -s -v tests/

Notes:

• The project is currently configured for local development and may require adjustments if you plan to deploy or run in ephemeral environments.
• If you intend to use the MCP inspector, you can install and run the inspector as described in the README, and then register the Data Wrangling server as shown in the example configuration.

Tips and considerations:

• When running locally, ensure that the path to your mcp-data-wrangler source directory is correct in the uv command (the --directory argument).
• If you plan to expose this as an MCP server, you may want to add additional environment variables or configuration options to control logging, data input formats, or result schemas.
• Since this server currently focuses on data wrangling tasks (data aggregation and descriptive statistics), you can implement or extend operations to include additional transformations as your workflows require.
• If you encounter issues with the inspector URL, verify that the local server is listening on the expected port and that network access between the inspector and the server is permitted.