Teradata ETL MCP Extension

Control your entire data pipeline from Copilot Chat: load data, generate dbt models, create Airflow DAGs, sync cloud apps, validate data—all through natural language.

Overview

Teradata ETL MCP Extension is a Model Context Protocol (MCP) extension that integrates data engineering tools into GitHub Copilot. Instead of switching between CLIs, UIs, and scripts, manage your entire ELT pipeline conversationally from VS Code.

What You Can Do

Ask Copilot to:

Load CSV files, sync cloud apps (Salesforce, Stripe, etc.), or copy tables between databases
Auto-generate dbt models from existing Teradata tables
Create Apache Airflow DAGs for scheduled pipelines and deploy them
Run SQL queries, execute DDL, validate data quality
Test dbt models and generate documentation
Monitor pipeline health and execution history

Everything runs against your Teradata instance. Credentials are secured and never exposed to chat.

Quick Start Video Guide

See the extension in action with these interactive demonstrations:

🔧 Setup & Configuration

Setup Teradata Credentials Initialize Teradata connections and configure the MCP server for your environment.

📤 CSV Data Loading

Load CSV Files Upload and transform CSV files into Teradata tables with intelligent schema detection.

🔄 dbt Transformations

Run dbt Projects Generate and execute dbt models for data transformation and testing.

🔀 Airbyte Integration

Airbyte Data Replication Create pipelines to replicate data from various sources (Postgres, MySQL, REST APIs, etc.) to Teradata.

⚙️ Airflow Orchestration

Airflow DAG Orchestration Orchestrate complex workflows combining Airbyte and dbt with Airflow scheduling and monitoring.

⌨️ Command Palette

MCP Server Commands Explore all available MCP tools and commands via the VS Code command palette.

Getting Started

1. Install

Search for Teradata ETL MCP in VS Code Extensions
Click Install

2. Setup (Built-in Wizard)

Open Command Palette (Ctrl+Shift+P)
Type: Teradata ETL MCP Extension: Setup Wizard
Fill in the form:
- Teradata: host, username, password, database, port
- Airflow (optional): base URL, username, password
- Airbyte (optional): base URL, client ID, client secret
Click Connect & Start
Extension registers the MCP server definition — start it from MCP: List Servers or by asking Copilot a question

3. Test

Open Copilot Chat (Ctrl+Shift+I) and ask:

"Check my Teradata connection"

Configuration

All configuration is handled through the Setup Wizard (built-in UI). Credentials are stored securely in VS Code.

Advanced: If you have existing config files (.env or connections.yaml), the wizard lets you optionally point to them via:

Command Palette → Teradata ETL MCP Extension: Setup Wizard
→ Advanced Settings tab
Provide file paths (optional)

Core Capabilities

1. Data Discovery

Understand your data without loading anything:

List tables — Query Teradata metadata (row counts, disk space, creation dates)
Inspect schemas — View columns, types, constraints, indexes
Profile data — Analyze data quality (NULLs, duplicates, cardinality)

2. Generate Code & Pipelines

Auto-create code from your data:

Generate dbt models — Scaffold SQL models from table metadata
Create Airflow DAGs — Generate Python code for CSV loading, table transfers, ETL workflows
Generate BTEQ scripts — Create Teradata utilities scripts

3. Load & Move Data

Multiple approaches for different scenarios:

Load CSV — Generate Airflow DAG (scheduled) or load directly via tdload (ad-hoc)
Sync cloud apps — Airbyte connectors (Salesforce, Stripe, Slack, etc.)
Copy tables — Teradata-to-Teradata transfers with optional dbt transforms
Export data — Extract to CSV, Parquet, or other formats

4. Transform & Test

Build and validate data models:

Run dbt models — Execute transformations, compile, build workflows
Test data — Run dbt tests (uniqueness, not-null, custom SQL)
Generate docs — Create dbt documentation

5. Execute & Deploy

Deploy pipelines and run operations:

Deploy Airflow DAGs — Ship DAGs to remote Airflow via SSH/SFTP
Trigger pipelines — Execute on-demand or on schedule
Execute SQL — Run queries, DDL, stored procedures directly
Run BTEQ — Execute Teradata BTEQ scripts

6. Validate & Monitor

Confirm success and track health:

Validate data — Run custom SQL checks, row count verification
Monitor DAGs — Check Airflow status, task history, runtime trends
Track tests — Monitor dbt test results over time
Pipeline health — Verify server connectivity, project validity

7. Manage Connections

Handle multiple environments:

List profiles — View all configured Teradata, Airflow, Airbyte connections
Reload credentials — Refresh connections after password rotation
Switch databases — Target different databases/environments per operation

System Requirements

Required

VS Code 1.100.0+
GitHub Copilot Chat extension
Teradata Vantage instance
Python 3.10+

Optional (Feature-Dependent)

Component	Used For	If Missing
Apache Airflow	Schedule/deploy pipelines	Generate DAG code (deploy manually) or use direct SQL tools
Airbyte	Cloud app syncing	Use CSV or direct SQL load instead
dbt	Transform & test data	Use raw SQL or Airflow instead
TdLoad (TTU)	Direct CSV loading (non-Airflow)	Use Airflow DAG approach instead
BTEQ (TTU)	BTEQ-specific scripts	Use teradatasql for SQL execution

Network Access

Connectivity to Teradata (host, port 1025)
SSH/SFTP to Airflow server (if deploying DAGs remotely)
API access to Airbyte (if using cloud connectors)

Typical Workflows

Fresh Database: Load → Discover → Generate → Run

User: "Load customers.csv and orders.csv to my database"
→ Copilot analyzes CSVs, generates Airflow DAGs, deploys, triggers loads
→ Creates STAGING.CUSTOMERS (50K rows) and STAGING.ORDERS (200K rows)

User: "Show me what tables I have"
→ Lists tables, schemas, row counts

User: "Create dbt models for both tables"
→ Generates dbt models + tests

User: "Run dbt and test the models"
→ Executes dbt run + dbt test
→ Reports: "✓ All tests passed"

User: "Schedule daily runs at 1am"
→ Creates Airflow DAG with schedule

Existing Database: Discover → Generate → Run

User: "List all tables in my ANALYTICS database"
→ Lists 15 tables with row counts

User: "Create a dbt model for CUSTOMERS"
→ Generates dbt model from table metadata

User: "Run the model and tests"
→ Executes dbt run + dbt test

Cloud Sync: Setup → Execute → Validate

User: "Sync Salesforce Accounts to my database daily"
→ Creates Airbyte connection, generates Airflow DAG

User: "Run the sync now"
→ Triggers Airbyte sync → Reports: "50,000 rows synced"

User: "Validate the data"
→ Profiles synced table → Reports data quality

Example Prompts (Copy & Paste)

Get Started:

"Check my Teradata connection"
"List all tables in my database"
"What's the schema for the CUSTOMERS table?"

Load & Move Data:

"Load sales_data.csv to my database"
"Create an Airflow DAG to load daily_sales.csv at 2am"
"Sync Salesforce Accounts to my database"
"Copy the ORDERS table from production to staging"

Generate Code:

"Create a dbt model for the CUSTOMERS table"
"Generate dbt models with tests for CUSTOMERS and ORDERS"
"Create an Airflow DAG for daily ETL"

Run & Test:

"Run dbt and show me the results"
"Execute dbt tests"
"Check if my Airflow DAG is valid"
"Validate that the load succeeded (row count check)"

Monitor:

"What's the status of my daily_refresh DAG?"
"Which of my dbt models is slowest?"
"Check my Airflow server health"
"Show me failed pipelines from the last 7 days"

Security & Privacy

Credentials secured — Passwords and API keys stored in VS Code SecretStorage; the Python MCP server resolves credentials from connections.yaml server-side
Response sanitization — Sensitive values masked automatically before appearing in chat
Audit logging — Optional trail of operations (no secrets exposed)
SSH security — Public key auth for remote Airflow deployment; host key verification supported

Pricing

Free to use. See the LICENSE file included with the extension for terms.

Support & Contributions

GitHub: Teradata-PE/devtools-elt-mcp-vscode-extension
Issues: Report bugs or request features
Discussions: Ask questions or share ideas

Troubleshooting

Connection Refused / Can't Connect to Teradata

Open Command Palette → Teradata ETL MCP Extension: Validate Connections
Verify Teradata host and port in Setup Wizard
Check firewall (default port: 1025)
Confirm credentials are saved in Setup Wizard
Ask Copilot: "Check my Teradata connection"

Server Not Starting

Open Output Channel: Teradata ETL MCP Extension
Check for Python/environment errors
Run: Command Palette → Teradata ETL MCP Extension: Recreate Python Environment
Ask: "Show Teradata ETL MCP Extension logs"

dbt Compile Failed

Verify dbt is installed: dbt --version
Ask: "Debug my dbt project"
Ask: "Validate my dbt project"

Airflow Not Responding

Verify Airflow server is running
Confirm base URL in Setup Wizard (local: http://localhost:8080)
Ask: "Check Airflow server status"

Configuration Issues

Open Setup Wizard: Command Palette → Teradata ETL MCP Extension: Setup Wizard
Review/update Teradata, Airflow, Airbyte settings
Click Connect & Start to apply changes
Run: Teradata ETL MCP Extension: Reload Configuration to re-register the server definition

CSV Load Hangs

Check file size (large files take time)
Verify Teradata has storage space
Check for table locks: RELEASE MLOAD <table_name>;
Try splitting into smaller batches

Additional Resources

dbt: docs.getdbt.com
Apache Airflow: airflow.apache.org
Airbyte: docs.airbyte.com
Teradata Vantage: docs.teradata.com
Model Context Protocol: modelcontextprotocol.io

Release Notes

v1.0.0

22 tools across 7 categories
Full Teradata, Airflow, Airbyte, and dbt integration
Secure credential management
Comprehensive error handling

Ready? Open Copilot Chat and ask: "List my Teradata tables"