ORATS – Driven By Data Ep. 143 — Safer Wheel Strategies
This episode is a genuine change of pace from the usual scanner-and-trade-setup format: instead of walking through a specific options strategy, Tyler demos ORATS’ new command-line interface, a tool built specifically to connect an AI coding agent like Claude directly to ORATS’ 15-plus years of historical options data. The pitch is straightforward but significant — rather than clicking through the dashboard, a trader can type a natural-language question into Claude, and the CLI hands the AI everything it needs to pull real ORATS data, respecting whatever data permissions (delayed, live, intraday) the trader’s account actually has.
Tyler kicks off the live demo by asking Claude to explain implied volatility conceptually and then evaluate which of SPY, QQQ, and IWM looks the most “stressed” right now — a genuinely good showcase of the tool’s dual strength, since it can teach a concept and immediately apply it to live data in the same response. Founder Matt Amberson, watching along, highlights a subtlety worth noting for anyone trying this themselves: there’s no single number that defines volatility stress, and a well-built agent should triangulate IV against its own history, against realized-volatility forecasts, and against the shape of the term structure and skew — exactly what Claude does in the demo, complete with a clean explanatory table.
The most genuinely useful segment covers Claude’s memory capability, which lets it retain a trader’s watchlist and preferences across sessions rather than starting from scratch every time. Tyler shows that Claude had already saved his watchlist from a prior session and, midway through this one, saves a new preference — using ORATS’ own historical volatility calculation (a modified Parkinson method, which Matt argues is meaningfully more accurate than the simplistic close-to-close method most tools default to) rather than the cruder standard. It’s a small moment, but it illustrates something valuable: the tool gets more personalized and more aligned with a trader’s actual methodology the more it’s used, without having to re-teach it every session.
From there, Tyler demonstrates spawning parallel sub-agents, each independently researching one ticker on the watchlist — pulling IV rank, term structure, and earnings risk — and compiling the results into a single written morning briefing document. Matt frames this as genuinely different from anything possible in ORATS’ existing dashboard AI assistant, since the CLI’s extensibility means a trader could, in principle, spin up hundreds of sub-agents to analyze an entire index at once. The segment closes with an attempt to schedule the whole workflow to run automatically each morning and email the results — which doesn’t fully succeed live on air due to an email authorization hiccup, but Tyler and Matt use the stumble constructively, walking through exactly how they’d troubleshoot it, which is arguably more instructive than a clean success would have been.
The episode wraps with practical setup guidance — where to get an ORATS API token, the different data tiers (delayed, live, live intraday going back to August 2020), and rough Claude subscription costs — positioning the CLI as a natural next step for traders already comfortable with Claude or a similar AI coding agent who want ORATS’ data working for them in the background rather than requiring a dashboard click for every question.
