Teradata MCP server in Flowise

📍 Navigation: Documentation Home | Server Guide | Getting started | Architecture | Installation | Configuration | Security | Customization | Client Guide

Make sure you have Teradata database access. (the most convenient way: Go to https://clearscape.teradata.com create account and login, start the environment and click on Run Demo)
Build Teradata mcp server container image from https://github.com/Teradata/teradata-mcp-server, run below lines in cmd terminal.

git clone https://github.com/Teradata/teradata-mcp-server.git
cd teradata-mcp-server
# build container from Source code
docker build --build-arg ENABLE_FS_MODULE=true \
   --build-arg ENABLE_TDML_MODULE=true \
   --build-arg ENABLE_TDVS_MODULE=true \
   -t teradata-mcp-server:latest .

Build Flowise Container Image from https://github.com/FlowiseAI/Flowise, run below lines in cmd terminal.

git clone https://github.com/FlowiseAI/Flowise.git
cd Flowise
docker build --no-cache -t flowise:latest .

Create Common .env file for teradata-mcp-server and flowise container,

mkdir ~/td_ai_stack
cd ~/td_ai_stack
vi .env

# ----------- MCP server and Database Env variables ------------#
DATABASE_URI=teradata://username:password@host:1025/databasename
LOGMECH=TD2  #TD2 or LDAP
TD_POOL_SIZE=5
TD_MAX_OVERFLOW=10
TDPOOL_TIMEOUT=30
PROFILE=dataScientist
DATABASE_HOST=IP_OF_DB_NODE

MCP_TRANSPORT=streamable-http  #stdio, sse, streamable-http
MCP_HOST=0.0.0.0
MCP_PORT=8001
MCP_PATH=/mcp/

# ----- Enterprise Vector Store ----------
TD_BASE_URL=https://host/api/accounts/40c83ff23b2e    #Your UES_URI, strip off the trailing /open-analytics
#TD_PAT=gwxhQG2UZcDqQlp9LKWjEBfXB7                     #Your PAT if you have Teradata Lake system.
TD_PEM=/root/td_ai_stack/demo_key.pem                 #Your PEM with full path where you kept on host                            
VS_NAME=vs_example                                    #Your target Vector Store Name

# ------------ Flowise env varieable -------------------#
PORT=3000
CORS_ORIGINS=*
IFRAME_ORIGINS=*
DATA_DIR=~/td_ai_stack/.flowise               # host dir to persist data of flowise

Create docker-compose.yaml file to up teradata-mcp-server and flowise containers

cd ~/td_ai_stack
vi  docker-compose.yaml

services:
    flowise:
        image: flowise:latest
        restart: always
        environment:
            - PORT=${PORT}
            # LOGGING
            - DEBUG=${DEBUG}
            # SETTINGS
            - CORS_ORIGINS=${CORS_ORIGINS}
            - IFRAME_ORIGINS=${IFRAME_ORIGINS}
           # Default Teradata Configuration env to refer into flowise
            - TD_MCP_SERVER=http://teradata-mcp-server:8001/mcp
        ports:
            - "${PORT}:${PORT}"
        extra_hosts:
            - "dbccop1:${DATABASE_HOST}"
        container_name: flowise
        healthcheck:
            test: ['CMD', 'curl', '-f', 'http://localhost:${PORT}/api/v1/ping']
            interval: 10s
            timeout: 5s
            retries: 5
            start_period: 30s
        volumes:
            - ${DATA_DIR}/.flowise:/root/.flowise
    teradata-mcp-server:
        image: teradata-mcp-server:latest
        restart: always
        environment:
            - DATABASE_URI=${DATABASE_URI}
            - LOGMECH=${LOGMECH}
            - MCP_TRANSPORT=${MCP_TRANSPORT}
            - MCP_PATH=${MCP_PATH}
            - MCP_HOST=${MCP_HOST}
            - MCP_PORT=${MCP_PORT}
            - PROFILE=${PROFILE}
            - TD_BASE_URL=${TD_BASE_URL}
            - TD_PAT=${TD_PAT}
            - TD_PEM=${TD_PEM}
            - VS_NAME=${VS_NAME}
        container_name: teradata-mcp-server
        extra_hosts:
            - "dbccop1:${DATABASE_HOST}"
        ports:
            - "${MCP_PORT}:${MCP_PORT}"
        volumes:
            - ${TD_PEM}:${TD_PEM}
        tty: true
networks:
  default:
    name: td-ai-stack
    external: false

Up teradata MCP server and flowise container

cd ~/td_ai_stack
mkdir ~/td_ai_stack/.flowise
docker image ls   # make sure teradata-mcp-server and flowise container images are available
docker compose up -d --remove-orphans

Validate docker container status

docker ps
# teradata-mcp-server container logs
docker logs teradata-mcp-server -f
# Flowise Container logs
docker logs flowise -f

Login to flowise http://IP:3000 or http://127.0.0.1:3000

first time login - Complete organization setup (set any username and password)

How to configure Teradata MCP server into Flowise Agentflow
- 9.1. Go Into Agentflows menu and Add new
- 9.2. Drag and Drop Agent and connect it with Start Node
- 9.3. Set up LLM credentials and LLM for Agent
  - Double click on Agent-0 , it will show Model to select model from various provider
  - select provider like Azure ChatOpenAI and set Azure ChatOpenAI Parameters,
  - for connect Credentails select -Create New-, fill details and Add
- 9.4. Add Teradata MCP server as custom MCP server for Tools
  - Click on Add Tools
  - Select Custom MCP server
  - Setup - Custom MCP Server Parameters
```
{
    "url": "http://teradata-mcp-server:8001/mcp",
    
}
```
  - Refresh Button of Available Actions
  - Click on Drop down to select tools for Agent
Save Agentflow with anyName
Execute AgentFlow