synthdata-serve

Expose synthetic datasets as a read-only MCP server.

Updated Apr 5, 2026
One-click install
npx skills add https://github.com/rappdw/synthdata --skill synthdata-serve
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: synthdata-serve
Source: https://github.com/rappdw/synthdata/tree/main/skills/synthdata-serve
Command: npx skills add https://github.com/rappdw/synthdata --skill synthdata-serve

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires numpy, pandas, openpyxl, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Spin up a read-only MCP server from synthetic datasets, enabling Claude to query and interact with tables without writing any server code.

Core Features & Use Cases

  • Expose data as an MCP server that Claude can query
  • Generate or load datasets (xlsx, csv, json, parquet) and auto-create query tools
  • Create an exportable standalone MCP server project for deployment

Quick Start

Launch a read-only MCP server from your dataset and start exploring tables with the built-in tools.

Frequently Asked Questions about synthdata-serve

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I expose a CSV or Excel dataset as a read-only MCP server for Claude?

You can expose CSV, xlsx, json, or parquet datasets as a read-only MCP server by loading the files into this Skill, which auto-creates query tools like list_tables and query_table for Claude to use.

Can I generate a standalone MCP server project from a parquet dataset for deployment?

Yes, you can generate a standalone MCP server project from a parquet dataset. This Skill supports exporting a self-contained server project that wires a read-only toolset for immediate deployment.

What data querying capabilities does an MCP server provide for Claude?

An MCP server for Claude provides data querying capabilities through a read-only toolset including list_tables, describe_table, query_table, sample_rows, and get_stats, allowing direct interaction with the loaded dataset.

Does this MCP server approach work with pandas and numpy data formats?

Yes, this MCP server approach works with pandas and numpy dependencies to load and process data formats like csv, json, xlsx, and parquet, wiring them into a read-only queryable interface.

What is the best way to make a synthetic dataset queryable by Claude without writing server code?

The best way to make a synthetic dataset queryable by Claude without writing server code is using this Skill to spin up a read-only MCP server, which automatically wires the necessary query tools for table interaction.