Wealth Management Agent Example using Pydantic AI Framework

This project demonstrates how to use Pydantic AI Framework with multiple agents working together. It leverages the supervisor pattern to "hand off" work to agents that have specific knowledge of a certain domain.

You will find a version of just using the Pydantic AI Framework and another version that leverages Temporal to wrap the agentic flow with Temporal.

The vanilla Pydantic AI framework version of this example is located here.

The Temporal version of this example is located here

Scenarios currently implmeneted include:

• Add Beneficiary - add a new beneficiary to your account
• List Beneficiaries - shows a list of beneficiaries and their relationship to the account owner
• Delete Beneficiary - delete a beneficiary from your account
• Open Investment Account - opens a new investment account
• List Investments - shows a list of accounts and their current balances
• Close Investment Account - closes an investment account

You can run through the scenarios with the Temporal version using a Web Application

Prerequisities

• uv - Python package and project manager
• [OpenAI API Key] (https://platform.openai.com/api-keys) - Your key to accessing OpenAI's LLM
• Temporal CLI - Local Temporal service
• Redis - Stores conversation history

Set up Python Environment

uv sync

Set up your OpenAI API Key

cp setoaikey.example setoaikey.sh
chmod +x setoaikey.sh

Now edit the setoaikey.sh file and paste in your OpenAI API Key. It should look something like this:

export OPENAI_API_KEY=sk-proj-....

Getting Started

See the Pydantic AI Framework version here And the Temporal version of this example here

Stay tuned! More to come!