Getting Started

Corrai is a strategy research & validation workstation — not a strategy vending machine. It helps researchers move from hypothesis, data, factors, and signals into validation, backtesting, and review, while making weak evidence and overfitting visible early.

The core idea

Most quant tooling optimizes for producing strategies. Corrai optimizes for judging them.

Every run flows through two stages:

1. Research DAG — creative freedom: draft hypotheses, build factors, compose signals.
2. Judge DAG — deliberately conservative: frozen artifacts, registered trials, statistical gates. It owns promotion. Approve or veto, with reasons attached.

The asymmetry is intentional. Corrai helps you produce more candidates — and is even better at killing them. What survives, survives with evidence.

Three ways in

Multi-agent research. A crew of research agents drafts hypotheses, builds factors, and runs experiments. Research memory carries lessons across sessions, so the same dead end is not explored twice. Agents propose; the Judge decides. See Agent Alpha Discovery for the full workflow.

Alpha Canvas. Compose research workflows visually — data, factors, signals, cross-sectional funnels — and run them reproducibly. Every run carries its lineage and comes back with a verdict. See Alpha Canvas Research Workflows.

Data Engine. Download market, on-chain, and alternative data into a local, governed lake: schema-standardized, version-fingerprinted, staleness-monitored. Evidence is only as good as its data. See Local-First Data Engine.

What the Judge checks

Every candidate faces the same gates:

• Purged CV + embargo — time-series-aware validation; random K-fold is refused
• DSR + PBO — Sharpe deflated by the number of registered trials
• Cost-aware execution — fees, slippage, next-bar fills, declared on every run
• Robustness — walk-forward, regime survival, cross-market stability
• Trial lineage — every experiment registered; the search history is part of the evidence
• Review & signoff — no researcher self-approval for production

If you want the reasoning behind these gates, start with Why Backtests Lie.

For a broader product overview, see AI Quant Research Workstation. For a validation-oriented checklist, see Evidence-Based Alpha Validation and Backtest Overfitting Checklist.

Early access

Corrai is in early access. There is no public pricing yet — request access and we will match you to the right tier. See plans for what each tier includes.