What is vibequant?

VibeQuant is a research tool for people who test trading ideas on financial markets, sometimes called quantitative research or quant research. It is built in Python on top of an existing engine called akquant, and it lets you start from a rough idea, a research paper, a forum link, or a PDF, and turn that into a structured plan the tool can actually run. Once you feed it an idea, VibeQuant can extract the concept using a language model, or fall back to simple keyword rules if no model is configured. From there you can run one of two kinds of research: factor research, where you test mathematical expressions against market data to see if they predict returns, or strategy research, where you backtest a full trading approach using built in templates. Data comes from free sources covering Chinese, Hong Kong, US, and crypto markets, with automatic fallback between providers if one is unavailable. A central part of the project is being honest about how trustworthy a result really is. Every backtest gets checked with a statistical test that gets stricter the more times you have already tested factors on the same data, plus a check for whether a pattern holds up across different time periods rather than just one lucky stretch. The project is explicit that this is not the same as true out of sample testing. Results can be reviewed as a bilingual report in English and Chinese, refined further by asking follow up questions, or handed to the tool to iterate on automatically. A validated strategy can also be deployed to send daily buy, sell, or hold signal emails after market close, though the tool never places trades itself. The README states plainly that VibeQuant is meant for research and education, and that nothing it produces should be treated as investment advice.