Stockpot
I wanted a stock tool that answered my questions without a terminal subscription or somebody else's scoring system.

Info
About this project
I needed a way to classify stocks quickly without paying for a terminal subscription. The backend selects the appropriate valuation family—from DCF and analyst blends to Owner Earnings, bank P/B, and REIT FFO—then builds a coherent conviction report, price-anchored expected return, scenario range, and investment thesis. The Angular client preserves those distinctions, shows evidence coverage and stale-data provenance, and never recomputes the finance in the browser.
- Implements niche valuation methods: Greenwald Owner Earnings and a custom Carrot Valuation
- Multi-pillar ConvictionReport: quantitative score built from valuation, quality, momentum, and data-integrity pillars
- Price-anchored expected-return architecture with separate strict margin-of-safety, action-entry, and scenario-hurdle prices
- 5-Factor Deal Quality model to score competitive advantages
- Reverse DCF calculator to surface the market's implied growth expectations for a ticker
- Valuation-family support for raw and guarded DCF, analyst blends, Owner Earnings, bank P/B, REIT FFO, and unavailable states
- Evidence-aware enrichments: analyst consensus, dilution filings, sanity-check coverage, DuPont ROE, Beneish M-Score, and normalized EPV
- Edge computing via Cloudflare Workers with KV caching for instant lookups
- Direct integration with SEC Edgar and Yahoo Finance for raw financial inputs


