Data Flow Generator · cwprep

Describe your data pipeline. Get a Tableau Prep file.

cwprep generates .tfl and .tflx files from natural language. No drag-and-drop in the Tableau Prep GUI required.

cwprep Tableau Prep flow automation preview

Quick Start

Install the package, configure MCP, then describe what you need. pip install cwprep

# MCP Server Config Example
{
  "mcpServers": {
    "cwprep": {
      "command": "uvx",
      "args": ["--from", "cwprep[mcp]", "cwprep-mcp"]
    }
  }
}

# Prompt Example
"Connect to intranet SQL Server (SSPI) to build a multi-database cleaning flow"
"Remove invalid columns, then execute Unpivot on month data"

Core Capabilities

SQL Output

See the SQL for every step

Every operation in your flow is translated to standard SQL.

Database Support

Your databases, supported

Native support for SQL Server and Alibaba Cloud MySQL.

No GUI Required

No Tableau Prep GUI needed

Generate complete .tflx packages without opening Tableau Prep.

All Operations

Common operations supported

Pivot, unpivot, join, union, filter, and calculate from plain text.

Capability Matrix

Feature	Datacooper	Default workflow	Advantage
Transparency	Full SQL output	GUI black box	Share SQL with DBA or compliance teams.
Runs without GUI	Prompt and code driven	Requires visual interface	Build many pipelines from scripts.
Your choice of AI	Any MCP-compatible model	Official cloud only	Use the model that fits your deployment boundary.

FAQ

Will my data be uploaded to the cloud?

No, unless you choose that setup yourself. cwprep can run locally through MCP, so data can stay on your machine or inside your own network.

Can I use cwprep without Tableau Prep installed?

Yes. cwprep can generate .tfl and .tflx files without Tableau Prep. You only need Tableau Prep to open and run the resulting flow.

What databases are supported?

cwprep supports SQL Server including Windows SSPI authentication, Alibaba Cloud MySQL, and standard ANSI SQL-oriented workflows.

Is cwprep bound to a specific LLM?

No. cwprep is model-agnostic and works with Claude, ChatGPT, Gemini, and DeepSeek when the model can drive MCP tool calls.

Core features open source. Advanced features and deployment on request.