Snowmelt is an AI earnings-options trading agent built on Alpaca. It scans the week’s US earnings calendar (~71 names), ranks reporters by market cap, and shortlists the top three per day into a nine-stock universe. From there it focuses on DELL, AVGO, and SNOW using three filters: historically large earnings moves (about 9.6–15.5% average absolute change), liquid OTM option chains with tight spreads, and shared AI-infrastructure theme risk. For each name, the agent builds an earnings prediction from fundamentals and guidance risk, then overlays quantitative OTM analysis (IV, spreads, volume, quote quality) to choose contracts and a directional lean—long calls when bullish, long puts when bearish—on the nearest post-earnings weekly expiry for max gamma. Risk is split across three OTM tiers (40%/30%/30% into 4–6%, 7–9%, and 10–12% OTM), sized to options buying power with safety buffers. Positions hard-exit by Sep 3 close so they are not held into expiry-day theta decay and IV crush; sells use live mid limit orders into the first RTH session after earnings until flat. On Alpaca, the agent pulls historical IEX bars and option chains for research, then orchestrates buys and sells through the Trading API with account, position, and market-clock checks.
Category tags: