AI Hedge Fund

AI Hedge Fund combines a set of signal generators — investor personas named for Buffett, Druckenmiller, Graham, Lynch and Munger, a post-earnings-announcement-drift signal, and a generic LLM agent — under four strategy mandates: deep value, earnings drift, fundamental long/short, and inflections. Around them the package ships a backtesting engine, portfolio construction, risk limits and event-study tooling. Mandates are YAML, so aihf can be driven interactively or non-interactively against a mandate file and a ticker list.

Nothing it produces reaches a market. The README states the project is for educational purposes only and that the system does not actually make any trades; the only broker implementation in the tree is a simulator. It is also not runnable on zero credentials: it needs a Financial Datasets API key for prices, fundamentals and earnings, plus a key for one supported LLM provider.

The project was restructured on 7 August 2026 and its distributable renamed to aihf (version 2.2.0, Python 3.11+), so descriptions written against the older package name or the previous src/main.py layout no longer apply — and the README announces a further rebuild toward a persistent always-on fund, so the interface is still moving. It is maintained by one individual rather than an organisation.