Teradata MCP Server

# Simple Agent - Using Google ADK framework (streamable-http version) This is a simple agent developed with the Google ADK framework that can use LLM. The ADK framework provides a simple user experience with adk web functionality. ## Features - Chat interface - Access to all tools - visibility into LLM messages - visibility into componenet execution time ## Prerequisites - Python 3.13 - LLM access (one of the following) - AWS - Account with Bedrock access - AWS CLI configured with appropriate credentials - Google - Google account with API key - Azure - Azure account with credentials - Azure model API key - Ollama - Ollama installed with MCP compliant model - Teradata MCP server and Teradata system. ## Installation 1. Install all client dependencies: With the server virtual environment activated, install the required packages: ```bash uv pip install -r examples/MCP_Client_Example/requirements.txt --force-reinstall ``` 2. Configure Client Credentials: Assumes you have set up the environment variables for your model. Alternatively you should add them to your .env file. ``` # When using AWS aws_role_switch=False aws_access_key_id= aws_secret_access_key= aws_session_token= aws_region= # When using OpenAI OPENAI_API_KEY= # When using Google GOOGLE_GENAI_USE_VERTEXAI=FALSE GOOGLE_API_KEY= # When using Azure azure_api_key= azure_gpt-4o-mini= # When using Ollama ollama_api_base= ``` 4. Modify agent.py file lines 55 - 78 to reflect the LLM model you are using. Default is usig AWS, to change, comment out the AWS model definition and uncomment the model you would like to use. <br><br> ## Usage 1. confirm the following is in .env file ``` MCP_TRANSPORT=streamable-http MCP_HOST=127.0.0.1 MCP_PORT=8001 MCP_PATH=/mcp/ ``` 2. In a termial start the server. ``` cd teradata-mcp-server uv run src/teradata_mcp_server/server.py ``` 3. In a second terminal start the client. ``` cd teradata-mcp-server source .venv/bin/activate cd examples adk web ``` 4. open [ADK Web Server ](http://0.0.0.0:8000) 5. Select the Simple_Agent 6. Chat with the agent