LLM agents that query live markets instead of making them up

What it does

openbb-agents is a Python package that turns natural-language financial questions into live data queries. It uses an LLM with function calling to interact with the OpenBB Platform, fetch up-to-date market data, and synthesize answers. You can ask for anything from a simple stock quote to a multi-step fundamentals analysis, and it will chain the necessary API calls together.

The interesting bit

The agent dynamically adjusts its available toolkit based on your configured data provider credentials, so it does not attempt to call APIs you cannot access. By default it falls back to yfinance for free data, but if you plug in premium OpenBB Hub credentials it will use those. It also handles temporal dependencies—queries where step two requires the result of step one, such as finding a company’s peers and then ranking them by market cap.

Key highlights

• Natural-language interface to the OpenBB Platform via LLM function calling.
• Supports complex, multi-step research queries with dependent sub-questions.
• Scopes available data sources dynamically based on your credential setup.
• Includes an async variant (aopenbb_agent) for non-blocking usage.
• Ships as a PyPI package, though it currently requires Python 3.10+.

Caveats

• The authors explicitly label it a work-in-progress.
• It is tightly bound to OpenAI; the README mentions no alternative LLM backends.
• Meaningful use beyond basic yfinance data requires separate OpenBB data provider API keys or an OpenBB Hub personal access token.

Verdict

Try it if you are already in the OpenBB ecosystem and want a quick way to prototype LLM-driven financial research. Look elsewhere if you need a production-grade, model-agnostic agent framework or if you lack the necessary financial data API credentials.