This repository demonstrates a financial analysis agent built with Agno, showcasing Urai’s powerful code execution capabilities.
This agent highlights how Urai’s code execution engine enables even lightweight language models like GPT-4o-mini to perform advanced financial analysis through code generation and execution.
For example, given a prompt such as Can you analyze SAP?, the agent dynamically generates Python code to retrieve financial data (e.g., from Yahoo Finance), then presents structured outputs like tables and charts.
Urai simplifies complex AI agent workflows by allowing language models to generate and run code in a secure, sandboxed environment. This decouples agent logic from model sophistication, enabling powerful outcomes at lower cost and higher speed.
Visit https://thiri.dev to sign up for an account and use the API key. Set the environment variable
THIRI_API_KEY to use the agent or the team of agents.
- uv
- Python 3.9+
git clone https://github.com/uraiai/financial_agent.git
cd financial_agent
uv venv
source .venv/bin/activate
uv pip install -r requirements.txt
python agent.pyGo to Agno playground URL. [https://app.agno.com/playground/agents]
Add an endpoint with URL http://localhost:7777 and NAME playground and click on SAVE.
Select the Financial Analysis Agent and type in something.
Eg: Can you give me an analysis on SAP?
This would give a graph and a neat summary of the competitive analysis.
View the agent.py that has the prompt and system instructions.
Note: You will need the Thiri API Key for code execution apart from your own OpenAI and Tavily API keys. Sign up at https://uraiai.com to get the API key.
Agno is a modular agent building library. Rather than imposing itself as the framework, Agno provides a set of tools that you can compose to build agents. This repo shows how Urai integrates seamlessly with Agno, enabling dynamic, code-driven agents that can take real actions.