Feature - add Teradata MCP server

[vilashjagani]

Added Teradata MCP server with its creds types , Teradata MCP server will be running Remotely

Add Teradata MCP serve for Agent

credential for Teradata MCP

TD docs - https://docs.teradata.com/r/Enterprise_IntelliFlex_Lake_VMware/ODBC-Driver-for-Teradata-User-Guide-20.00/Network-Security/Authentication-Mechanisms

** Chatflow MCP server**

[vilashjagani]

LGTM

[0xi4o]

Hi @vilashjagani. Thank you for your contribution.

Can you add links to documentation for:

1. Where to get credentials for both bearer token and TD2?
2. Where to get the MCP Server URL? And if possible, how to set it up locally?

[vilashjagani]

Hi @0xi4o -
1. Where to get credentials for both bearer token and TD2?

https://docs.teradata.com/r/Enterprise_IntelliFlex_Lake_VMware/ODBC-Driver-for-Teradata-User-Guide-20.00/Network-Security/Authentication-Mechanisms

Teradata users know what TD2 creds - its username and password to connect Teradata DB system
to generate simple basic auth using Teradata DB username and password

import base64

username = input("Enter username: ")
password = input("Enter password: ")

# Combine username and password with colon
user_pass = f"{username}:{password}"

# Encode to base64
basic_auth_creds = base64.b64encode(user_pass.encode()).decode()

print(f"Basic Auth Credentials: {basic_auth_creds}")

same way if Teradata DB is integrated with any IDP provider solution to authenticate based access token issue from IDP -
simple example to generate JWT from IDP provider

 # This will Generate token from IDP
    token_gen_url = f"{auth-idp-provider-fqdn}/protocol/openid-connect/token"
    payload = {
            "grant_type": "password",
            "client_id": client_id,
            "username": username,
            "password": user_password
        }
    
    response = requests.post(token_gen_url , data=payload)

if IDP support Device flow to gen token

device_auth_url = f"{{auth-idp-provider-fqdn}}/protocol/openid-connect/auth/device"
    token_url = f"{{auth-idp-provider-fqdn}/protocol/openid-connect/token"
    response = requests.post(device_auth_url, data={
        "client_id": client_id,
        "scope": "openid"
    })
    
    data = response.json()

  # Show user-friendly URL and Token
    display(print(f"""
    ### 🔐 Authenticate Using Device Flow To Get Access Token - 
    [🌐 Click on Link]({data['verification_uri_complete']})  
    or open `{data['verification_uri']}` and enter this code:  
    # `{data['user_code']}`
    """))

2. Where to get the MCP Server URL? And if possible, how to set it up locally?
this teradata MCP server will be hosted into Teradata account with Teradata Database system so customer have teradata-mcp server URL into their account of teradata

Setup locally - will be published as public facing docs in short time with more details or can be container image itself to run it.

[HenryHengZJ]

thank you!

[HenryHengZJ]

As we can't get access to Teradata by just signing up, do you have a couple of screenshots that show how would this integration looks like in actual use case?

After this get merged, I think it would be great if docs can be added as well - https://github.com/FlowiseAI/FlowiseDocs

[HenryHengZJ]

As we can't get access to Teradata by just signing up, do you have a couple of screenshots that show how would this integration looks like in actual use case?

After this get merged, I think it would be great if docs can be added as well - https://github.com/FlowiseAI/FlowiseDocs

@vilashjagani do you have working screenshots or recordings (better) to showcase the Agent using Teradata MCP?

[vilashjagani]

Hi @HenryHengZJ
TD2 Auth Setting for TD MCP server -

T56nUS5sZ6.mp4

JWT based Auth Cred settings for TD MCP server -

GPgOSGouCd.mp4

[vilashjagani]

Hi @HenryHengZJ - Can you please have look of recording How teradata MCP server can be use and merge this PR if everything is fine.

[HenryHengZJ]

thanks for the effort @vilashjagani !

[HenryHengZJ]

next step: can you update the docs as well? https://github.com/FlowiseAI/FlowiseDocs