Institutional-grade trading research produced by a team of autonomous AI agents. Hypothesis-driven. Backtested. Falsified. Delivered weekly.
Every report follows a rigorous process: hypothesis, backtest, falsification, verdict. We reject more ideas than we publish.
Original trading ideas across perpetual futures, funding rates, liquidation cascades, basis trades, on-chain flow, and options skew. Not recycled TA.
Every idea is tested across market regimes with realistic cost assumptions. In-sample / out-of-sample splits. Walk-forward validation. Parameter sensitivity.
A dedicated review agent tries to kill every idea before publication. Does it survive costs? Is it just beta? Is it overfit? Only robust ideas ship.
15-minute market signals across BTC, ETH, SOL, XRP, DOGE. Momentum, mean-reversion, fade, and dip-buy signals with entry/exit levels.
Automated CoinDesk monitoring filters noise into actionable context. Know what matters before the market prices it in.
Every idea gets a verdict: Reject, Watchlist, Promising, or Ready. No ambiguity. No hedge-everything disclaimers. Honest assessments.
Each weekly report contains 1-2 fully developed research ideas plus a market signal summary.
Edge survives transaction costs. Robust across regimes. Degrades at high AUM due to capacity constraints. Recommended for further live testing at small size.
Hypothesis, mechanism, signal definition, trade construction, data sources, test design, and results.
Where it breaks, what invalidates it, whether it's just beta in disguise, and specific kill criteria.
All MOMO, FADE, DIP-LONG, and PROBE signals generated across BTC, ETH, SOL, XRP, and DOGE.
Curated news and macro context that drove the week's price action and informed our research direction.
Ideas in the pipeline, signals to monitor, and setups we're tracking for the coming week.
Our research is produced by a team of specialized AI agents orchestrated through Paperclip, with human oversight on every published report.
Coordinates all agent activity, assigns research priorities, enforces quality gates.
Primary research engine. Discovers, tests, and validates novel trading hypotheses.
Adversarial reviewer. Stress-tests every idea and tries to falsify results before publication.
Builds and maintains data pipelines, backtesting infrastructure, and signal generation systems.
Core system architecture. Designs execution flow, API integrations, and platform reliability.
Final review on every report. Sets research direction. Makes the call on what ships.
No tiers. No upsells. Full access to every report, signal, and research archive.
Cancel anytime. No commitment.