An autonomous AI platform that screens Indian equities daily using parallel agent signals, a 6-gate fundamental filter, and Claude AI-generated investment memos.
Parallel AI Agents
Evaluation Gates
Memo Sections
Core Use Cases
Upload 52-week high files via web UI or API, or auto-fetch from BSE India. Fundamentals and news pulled from Screener.in, Yahoo Finance, and BSE feeds.
Three parallel AI agents evaluate price proximity, revenue growth, and news sentiment simultaneously with weighted confidence scoring per ticker.
Every stock passes through six configurable evaluation gates: PE ratio, market cap, sales growth, sentiment, management quality, and negative news ratio.
Claude AI generates comprehensive investment memos with 20+ sections per stock covering thesis, fundamentals, risk, and valuation. Export as PDF, DOCX, or ODT.
Live SSE-streamed pipeline monitoring, one-click runs, KPI dashboard, upload triggers, and full run history. Edit all thresholds from UI - no code changes.
Role-based access: Reader, Writer, Admin. Per-user data isolation for uploads, runs, watchlists, and reports. Login audit logging with full run traceability.
Each agent runs independently and contributes a weighted confidence score to the final merged signal.
52-week high proximity. Default threshold >= 95%.
YoY revenue & PAT growth. Quarterly results analysis.
News quality & sentiment. BSE / Yahoo News feeds.
Per-stock outcome: SELECTED • REJECTED • SKIPPED
<= configurable threshold
Small-cap focus filter
>= minimum YoY %
Positive score >= threshold
Score >= minimum
<= capped threshold
Ten distinct workflows covering analysts, schedulers, admins, and readers - from daily screening to disaster recovery.
Upload 52-week high CSV -> trigger pipeline -> view Diamond Watchlist with picks instantly.
System fetches external data at configured time, runs all agents, generates picks automatically.
Open Diamond Report to read Claude AI-generated investment memo covering thesis, fundamentals and risk.
Change PE cap, growth thresholds, or sentiment weights from Settings UI - no code changes needed.
Run backfill with a date range to regenerate signals for past periods and validate strategy performance.
Check /api/source-health or dashboard tiles for stale, broken, or unavailable data providers.
Download screened signal outputs as CSV or JSON from the exports API for spreadsheets or downstream tools.
Browse History page to review which stocks were selected, rejected, or skipped on any prior date.
Each user uploads their own data and views only their pipeline results, watchlists, and reports.
Admins run backup-db and restore-db CLI commands to snapshot the database or rollback safely.
Configurable hour/minute refresh - runs pipeline unattended on schedule.
One-click pipeline run from dashboard UI or REST API call.
Historical date range processing to regenerate past signals on demand.
Auto or manual trigger to pull fresh data from BSE, Screener, Yahoo.
CLI-driven snapshot and rollback
Signal export via exports API
HTML archival of run states
System and source health monitoring
/api/metrics for run telemetry
/api/alerts for threshold breaches
Full run IDs for reproducibility
No hardcoded business thresholds
See how Diamond Picks can transform your equity research with AI-powered screening and investment memos.