The BEST Way to Build Intelligent Apps with Phidata AI Agents

---

---

AI Summary

Summary of Video Transcript: Creating AI Agents with FI Data

Overview

• The video tutorial covers the creation of AI agents using FI Data, an open-source tool.
• It is designed for beginners to create AI agents capable of complex tasks, automation, and decision-making.
• The code for the tutorial is available in the FI Data repository under “cookbook/agent11”.

AI Agents

• AI agents are independent AI systems that work together to solve complex tasks.
• The tutorial demonstrates building a web search agent and a financial analyst agent.
• These agents will be combined to perform tasks like searching for information on Nvidia and Tesla and checking their stock prices.

User Interface (UI)

• A UI is set up to interact with the created agents.
• The UI allows users to assign tasks to the agents and receive combined results from their tools.

Steps in the Tutorial

1. Setting Up the Environment
   • Install necessary packages using pip install.
   • Export the OpenAI API key from platform.openai.com.
2. Importing Packages
   • Create a file app.py.
   • Import packages from FI Data, including agent, openai, chat, dugdug go, Yahoo finance tool, SQL agent, storage, and serve playground app.
3. Creating AI Agents with Tools and Memory
   • Define the web search agent with dugdug go search for internet searches.
   • Define the financial analyst agent with Yahoo finance for stock prices.
   • Both agents are given instructions and set to store data locally with markdown formatting.
4. Creating the User Interface
   • Set up the UI with the serve playground app.
   • Run the code in the terminal using fy commands to authorize the UI and start the application.
   • Access the UI through a provided local URL.

Testing Agents

• Individual agents can be tested through the UI.
• An agent team can be tested by asking it to perform a combined task, such as searching for information on Nvidia and Tesla and finding their stock prices.

Conclusion

• The tutorial provides a foundation for building and testing AI agents with FI Data.
• Additional tutorials on FI Data are available and recommended for further learning.