etl-pipeline-mcp 1.0.0

A sovereign, local-first MCP-Native ETL System with Hot/Cold tiering, PII scrubbing, ID stitching, and warehouse exports.

seosiri-etl-pipeline-mcp

A sovereign, local-first, open-source Model Context Protocol (MCP) server for end-to-end Big Data Extraction, Transformation, Identity Resolution, and Warehouse Loading (ETL/ELT).

💖 Sponsorship & B2B Custom Solutions

👨‍💻 Lead Architect & Attribution

Designed and engineered by Momenul Ahmad, Lead Architect and Founder of SEOSiri.

🚀 Capabilities

• Multi-Platform Ingestion: Ingests real-time webhooks (CMS, email, social) and batch CRM records (HubSpot, Salesforce).
• Decoupled Hot/Cold Storage: In-memory SQLite Hot Tier for low-latency queueing + on-disk Cold Tier for permanent, anonymized storage.
• Identity Resolution: Merges fragmented user identities across CRM, social, and email records under a single mcp_root_id via ID stitching.
• GDPR/HIPAA PII Redaction: Automatic SHA-256 hashing and string redaction of PII before long-term storage or export.
• Warehouse Exporter: Exports clean, transformed data buffers directly to Snowflake, ClickHouse, or BigQuery.

Quickstart

1. Install in Editable Mode:

   pip install -e .
   

2. Run Unit Tests:

   pytest tests/test_etl_pipeline.py
   

🔌 Connecting to Claude Desktop / Cursor

Local Script Execution

Add this block to your claude_desktop_config.json:

{
  "mcpServers": {
    "etl-pipeline-mcp": {
      "command": "python",
      "args": [
        "D:/seosiri-etl-pipeline-mcp/src/main_server.py"
      ],
      "env": {
        "PYTHONPATH": "D:/seosiri-etl-pipeline-mcp"
      }
    }
  }
}

Remote Execution via GitHub

Or for zero-setup execution directly from GitHub using uv:

{
  "mcpServers": {
    "etl-pipeline-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--github",
        "SEOSiri-Official/etl-pipeline-mcp",
        "src/main_server.py"
      ]
    }
  }
}

License

Distributed under the MIT License. See LICENSE for more information.