🎯 Decisions

# AI-Powered Web Design Process *Strategy for WiderWings products (MedSchools.ai, CollegeDojo, Hedge, etc.)* Created: 2026-03-02 Source: Two YouTube videos on AI-based web design (transcripts in Seco

Henry corrected the data provider hierarchy. Schwab is PRIMARY for ALL data types (daily AND intraday), not just daily. Correct order: 1. Schwab (primary, if authenticated) 2. Alpaca (secondary fallba

Daily: Schwab→yfinance. Intraday (1m/5m/15m/1h): Alpaca→yfinance. Unified provider at data_provider.py. Opening Power updated to use this. Henry confirmed this hierarchy for ALL new strategies.

Hedge Product Strategy approved 2026-02-21. Three Pillars: 1. Research (Charts, Screener, News, Pattern Scanner) - find opportunities 2. Strategies (Regime Lab) - build regime-driven allocation model

## Context Henry wants to turn me (Mark Wings) into a world-class scraper. Three layers: (1) Research & Recon — find where data lives, (2) Scraping & Extraction — pull data at scale, (3) Data Structur

## Decision Chose Python FastAPI for Hedge backend instead of keeping everything in SvelteKit. ## Reasoning 1. **ML/Quant Libraries**: pandas, ta-lib, scikit-learn work natively in Python 2. **Backte

## Tools Evaluated ### Nano Banana (Google Gemini) - Gemini native image generation - Part of Gemini API (cheap at scale) - Good for: text-to-image, photo editing, character consistency - Best long-t

## Pricing Structure (Approved 2026-02-06) ### Base Plans - **Free**: $0 - Full platform except Interview, 5 AI chats/day - **Monthly**: $49.95/mo - $1 for 3-day trial, then full price - **Applicatio

## Context After setting up Second Brain, Bob failed to save email auth work during the conversation. Henry called this out. ## Decision **Save too much, not too little.** The cost of saving is cheap

## Question How frequently should agents write to Second Brain? Periodic? Always? Triggered? ## Decision **Trigger-based saves** — Save when something meaningful happens, not on a schedule. ## Reaso