VIRGO
Your ideas. Working strategies.
When I first explored quantitative investing, I wanted to build investment strategies by talking with AI. But writing code, gathering the material I needed, checking for errors and analysing results were separate tasks. I built VIRGO for my own use, to bring that process together in one place. I hope beginners can start by talking through an idea, while experienced users can work freely in their own way.
Describe an investment idea, build the code and material with AI, and refine it through inspection and analysis. A quant workspace for beginners and experienced developers to work in their own way.
9 min 15 s · Build, inspect, allocate and review in the actual app. VALE conversations demonstrate a workflow; performance and allocation examples use historical data and scenario simulations.
From building a strategy to understanding its results.
Describe it. See it as cards.
Arrange code, charts, tables and material around your work.
Inspect and refine together
Review technical errors and risks, then revise with VALE.
Allocate and learn from results
Define allocation rules and review the results in reports.
A strategy room for each investment idea.
Quant investing starts by turning an idea into rules you can check against data. Translate an idea such as buying when price is above its long-term average into code, then examine how it would have behaved in past markets.
A strategy room brings together code, stock and trading-frequency settings, documents, graphs and inspection results. Develop ideas such as trends, pullbacks, breakouts and volatility responses in separate rooms and compare them.
Turn a request into code and cards.
“Build a trend strategy for MSFT and JPM. Show price charts, moving averages, a comparison table and performance graphs.”
VALE is a separate AI workspace that connects models such as GPT, Claude and Grok. Connect it to VIRGO to discuss the strategy, edit its code and create the cards you need alongside your work.
Keep the VALE conversation in a separate window and use the VIRGO workspace to view the results. Experienced programmers can edit directly and ask for help where they need it.
Connect to VALE from another device →Put the analysis you need where you need it.
See prices, moving averages, stock comparisons, equity curves, drawdowns and monthly returns together as cards. Add analysis for the question at hand, collapse code or expand a chart, and shape your own workspace.
Move and resize cards yourself, or describe the layout you want to VALE.
Find easily missed errors before use.
Inspect execution, trading signals, costs, performance and overlap between strategies. Technical failures block approval, and the results show which checks failed and why.
Select Ask VALE to bring the results into a conversation, investigate the cause and revise the code. Beginners get a starting point, while experienced users can catch issues hidden in complex rules.
Stress testing is part of the inspection. It examines higher costs and real market shocks, including the Lehman crisis and COVID crash, to reveal vulnerable periods. Risk warnings highlight issues to review separately from technical failures.
Allocation is a strategy in its own right.
The allocator holds the rules for running strategies together. Examine the combination to avoid concentrating capital in strategies that move alike, and define how much each receives under different conditions.
In the video, higher market volatility reduces total exposure. Later, when the market stabilises but the breakout strategy alone underperforms, its weight falls from 10% to 2%. Other weights stay in place, the released capital moves to cash, and the reason is recorded.
Inspect the allocator’s code and conditions too, then analyse and refine the results with VALE.
Start with code you bring in.
Review files and their source on GitHub, then import them into a strategy room. Ask VALE to explain the code, adapt it to your conditions and VIRGO, then follow the same execution and inspection process.
Read the results. Refine what comes next.
Review returns alongside maximum drawdown, volatility, the Sharpe ratio, a benchmark and monthly results. Ask VALE about unfamiliar measures and use difficult periods as evidence for your next revision.
Export metrics and records to Excel for comparison and analysis, and save reports as PDFs.
Connect an account through the Alpaca API.
Register your Alpaca API keys in Settings to connect accounts and orders. Check the order workflow with a paper account first, then connect a live account for actual operation.
Automated trading stays off until you enable it. Strategies enter operation after inspection, the required review and your approval.
VIRGO runs on your computer, with broker keys stored in the local keychain.
Describe an idea, build it, inspect it and learn from the results. You can develop your own strategy step by step as you work.
Download
VIRGO runs on your own computer. The Mac installer is signed and notarized by Apple; the app tells you when a new version is out.
View releases on GitHub ↗