Traditional options discipline.
Agentic AI conviction.
One dashboard.

It is a full options desk in one dashboard. A morning scanner ranks your universe by signal strength and IV rank. A conviction engine scores every ticker with the drivers shown. A trade plan builder turns any signal into strikes, expiry, probability of profit, and a breakeven cone. A live monitor watches every open position and flags hold, roll, take profit, or close. A journal auto-classifies every trade and tracks what actually works for you. And an AI assistant sits on top of all of it to explain anything on screen in plain English.

The rules. Every signal, entry filter, and risk cap comes from a deterministic rules engine built on long-standing options industry practice: defined-risk structures only, 30 to 45 days to expiry, short strikes inside a disciplined delta band, IV rank floors so premium is only sold when it is rich, liquidity minimums on open interest and bid-ask width, credit-to-width floors, and position sizing computed from worst-case loss. The same inputs always produce the same outputs. The AI cannot override, loosen, or talk its way past a single one of them.

In two places, both deliberately bounded. Inside the engine it is secondary guidance: an AI perspective that reviews what the rules already scored, adds market context, and never changes the math. Where it truly shines is the Help & AI section: a research assistant with more than twenty live tools that reads your actual positions, explains every measure and factor in the app, and answers account-specific questions like "what is my riskiest position this week" or "why was that spread rejected."

The auto trading agent now manages the book, not just the trade. Every position is beta-weighted against SPY, and three deterministic gates keep the portfolio balanced: directional exposure must stay proportional to daily theta income, is hard-capped as a fraction of capital, and when the book tilts too far, the agent will only accept trades that pull it back toward equilibrium. Every rejected candidate is logged with its exact reason, and a built-in AI analyst explains the aggregate picture in plain English.

Grounding is enforced, not promised. Every number the assistant states must come from a live tool call against your account or market data inside that same conversation, and it cites the tool it used inline. The server independently checks each reply against the tools that actually ran. If the assistant cannot ground an answer, it says "I need to run the tools for that" instead of guessing.

Ask why your conviction score is an 87 and you get the real subscores that built it. Ask why the engine rejected a trade idea and you get the exact filter that failed and the number it failed by, read from the record the engine wrote at decision time. Nothing is reconstructed after the fact and no rationale is invented to sound plausible. Every measure, factor, and account-specific decision in the app can be traced to its logged reasons.

Not with real money without you. A research agent can paper-trade the rules discipline in a sandbox so you can watch the process work risk-free. Anything touching a real account requires your explicit review and approval on every single order. The AI can analyze, propose, and queue an idea for your review. Only you can execute.

Yours. You connect your own Schwab and or Tradier account with a one-time OAuth login and the platform works from your broker's real-time feed. Your credentials never touch our servers, nothing is resold, and you never pay twice for data you already have.

No. Options Scanner is a personal project and an analysis tool, not a registered investment advisor. Signals and scores come from rule-based and statistical analysis of public market data plus your own account information. Every trade is your decision and your responsibility.