Huginai is a one-person crypto research desk. We ran a signal-driven directional book for three cohorts, measured a 40% win rate against the 57.4% it needed to break even, and closed it. The one strategy still running is a funding-carry test on Hyperliquid, on paper.
The pipeline is real and it runs. What it produced is what closed the book — the ingestion works fine, the entries had no edge.
Public Crypto Twitter, Telegram channels and on-chain flows land in one stream. This part works and still runs.
Related mentions are grouped and de-duplicated, sources weighted, and each cluster scored 1–10 with the sources kept. The score turned out to be anti-predictive across all three cohorts, so it no longer gates or sizes anything.
Everything is traded on paper against live prices, with taker fees and slippage modelled. No real capital, ever.
A cohort either clears its breakeven win rate or it is closed. Three did not. One benchmark that looked good turned out to be a daemon outage holding two positions through a rally, and was thrown out too.
Short the Hyperliquid perpetual, hold the spot leg, collect the funding difference. Delta-neutral, so it does not need a directional view — which is the part that failed. Measured over 60 days on paper, with taker fees charged on both legs.
Not a live track record, and not an offer. This is a paper forward-test on a single operator's own account, with a kill rule that closes a leg if the funding floor drops below 7bps over seven days. Single-digit APR on a delta-neutral carry is a boring, believable number — which is rather the point. Anyone quoting you three digits is selling something.
This is one person's research project, not a product. There are no plans, no accounts and no checkout — the pricing page that used to sit here described a purchase path that was never built.
The desk interface is published as a demo build. It renders synthetic sample data behind a permanent banner saying so — no account, no live market data, no positions behind it. It exists so you can see how the thing is put together.
Open the demoThe verdicts, the autopsy on the directional book and the carry forward-test methodology are written up in the repository, including the parts that did not work. The blog covers the same ground in longer form.
Read the write-upsNo capital is managed for anyone. Nothing here is financial advice, and every figure on this site is a paper measurement of a single operator's own experiments.