Finny is an AI-powered trading assistant designed for algorithmic investing, offering both a command-line interface (CLI) and a web application. The platform enables traders and developers to move from describing a trading strategy in plain English to generating inspectable Python code, conducting backtests with historical data, and progressing to paper trading or live trading using their own execution systems. Finny is positioned for users who want to build, validate, and iterate on algorithmic trading strategies, emphasizing transparency and reviewability at each step.
Key features include the ability to describe strategies in natural language, automated research through subagents that gather data and news, and the generation of Python code that can be inspected and reviewed. Backtesting is supported with eight metrics on historical data, and users can paper trade to test strategies with zero risk before moving to live trading. The platform allows integration with various models, including OpenAI, Anthropic, DeepSeek, and custom financial models, by using user-supplied API keys. Finny also supports running its own self-hosted models. The tool is designed to work with existing trading infrastructure, such as data feeds, backtesters, and brokerage workflows, including compatibility with Oracle DB, Azure Data Lake, Bloomberg Terminal data, and IBKR brokerage.
Finny provides two main interfaces: a terminal-based CLI and TUI for builders to describe, inspect, and run strategies locally, and a browser-based web app for teams to review, compare, and deploy strategies with audit trails and deployment gates. The system emphasizes security and compliance, offering sandboxed runs, bias and safety checks, and evidence packets for sign-off. Users can install Finny locally via npm, and there are options for deeper integration with existing trading workflows through consultation.
The platform offers different product tiers, including a Pro version with unlimited local runs, backtests, and workflows, as well as a cloud-hosted version for live runs, status automation, and managed infrastructure. Finny is not a budgeting app or robo-advisor; it is tailored specifically for algorithmic trading strategy development and deployment.
In the Autonomous agents & workflows space, Finny Ai takes a focused approach. It focuses on automating the creation, validation, and backtesting of algorithmic trading strategies from the command line. Finny Ai is a B2B product aimed at algorithmic traders and developers. A free plan is available. Finny Ai is available on the command line.
Finny builds and maintains Finny Ai, and it first shipped in 2025. The project is developed in the open on GitHub with 49 stars and 796 commits in the last 90 days. Among its 5 catalogued features are strategy generation, backtesting, and validation.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do