30 red-flag checks.
Under 5 minutes.
Per ticker.
D30 reads the annual report, exchange announcements, and live market data - then runs a 30-question forensic framework across Strategy, Operations, Financial, Reputation, and Investor Relations. The output is a plain-English CLEAR / REVIEW / AVOID verdict plus the Earnings Integrity Score (P/E ÷ Q).
Built by Berkeley Advisory. Multi-org from day one. Every scan persisted as structured JSON for future fine-tuning.
30 questions · 5 categories
Strategy, Operations, Financial, Reputation, Investor Relations. Each question resolved by 5 - 9 forensic sub-checks against the AR, exchange announcements, and live market data.
Earnings Integrity Score
Q = count of GREEN-verified questions. EIS = P/E ÷ Q. Cheap per unit of verified quality. Plots live on the four-quadrant Quality-PE matrix.
Training-data from day one
Every scan persisted as structured JSON with prompt version + metadata snapshot. Outcome field filled when real-world events occur. A proprietary dataset no competitor can replicate.