Cloud Foundry sample application that provides chat interface for interacting with Goose AI agent with Tanzu AI Services

Products

• VMware Tanzu Platform - Cloud Foundry

Issue/Introduction

How to create a full-stack web application in Cloud Foundry for providing a chat interface for interacting with Goose AI agent. 

Customer is using Tanzu AI Services and would like to integrate with chat interface.

Sample application: https://github.com/cpage-pivotal/goose-agent-chat

Environment

Tanzu AI Services

Resolution

Download the sample goose-agent-chat web application.

This getting started guide details how to successfully integrate application with Tanzu AI Services.

Follow the needed steps to configure application and AI service configuration - 

• Configuration Overview
• Configuring LLM Providers
• Adding MCP Servers
• Credential Management with the Agent Credential Broker
• Configuring Skills
• Building and Deploying
• Cloud Foundry Deployment
• Tanzu Marketplace Integration

Configuring LLM Providers

There are two ways to configure which LLM model to use:

1. Tanzu GenAI Service Binding (recommended for Cloud Foundry) - Bind a GenAI service from Tanzu Marketplace, and models are automatically discovered
2. Manual Configuration - Set provider and model in .goose-config.yml with API keys

Note: GenAI service bindings take precedence over manual configuration. See Tanzu Marketplace Integration for details.

Here's a sample manifest for using sample app with the (Tanzu AI) Agent Buildpack (if 10.4.x) -

applications:
- name: my-agent
  memory: 1024M
  buildpacks:
  - agent_buildpack
  services:
  - your-llm

Binding a GenAI Service and Restage app

# Create a GenAI service instance from Tanzu Marketplace
cf create-service genai standard my-genai-service

# Bind to your application
cf bind-service goose-agent-chat my-genai-service

# Restage to pick up the binding
cf restage goose-agent-chat

Reference Tanzu AI Service Docs: Deploy an AI Agent.