FinRobot
An Apache 2.0 financial AI agent app with local valuation computation, cloud or custom local model connections, and a separate self-hosted report service.
Alternative to:
FinRobot is an Apache 2.0 open-source equity research system for analysts and quantitative researchers. Its desktop app runs financial calculations locally and uses an LLM to interpret results and write reports. Built-in OpenAI and Anthropic connections send model requests to those providers; custom OpenAI-compatible connections can use a local vLLM backend. Research also needs external market data.
The native desktop app runs on macOS with Apple Silicon. The same engine has a local browser interface for Intel Macs, Windows and Linux, plus a command-line interface. Existing reports remain accessible offline.
Python code calculates valuations, including discounted cash flow, dividend discount models, leveraged buyout analysis, peer comparisons and Monte Carlo simulations. Figures retain their calculation provenance, and audit checks flag disagreements between the numbers and the generated narrative.
Research workflows cover company analysis, earnings reviews and investment committee memos. Specialist agents gather data and draft reports, while bull and bear agents present opposing cases for a judge agent. The stock workspace pairs live market data with valuations; reports include target ranges, confidence assessments and data caveats. Data sources include yfinance, SEC EDGAR, FMP and Finnhub.
FinRobot Equity is a separate self-hosted browser service that produces HTML and PDF reports. It requires OpenAI and FMP API credentials. The hosted service at finrobot.ai is distinct from these user-run applications, and the AutoGen library provides educational financial-agent examples.
